Name: dev-init
Author: rhixecompany

Prompt file imported from rhixecompany/sandbox (.github/prompts/development/dev-init/dev-init.prompt.md). Copyright stays with the author.

name: dev-init title: Development Initialization description: Initialize a project for prompt-driven development: discover, plan conversion, audit/fix prompts, execute an implementation plan, and verify. trigger: /dev-init category: development version: 1.0.0 author: Hermes Agent tags: metadata: hermes: profile: code-architect priority: medium copilot: model_required: sonnet opencode: enabled: true codex: toolsets: skills:

skill: using-superpowers dependencies: [] formatter: markdown license: MIT

Goal

Initialize a project for prompt-driven development: discover, plan conversion, audit/fix prompts, execute an implementation plan, and verify.

Context

Phases

Goal
Context
Inputs
Outputs
Rules
Skills Required
Phases
- Phase 1: Discover and verify
- Phase 2: Plan conversion
- Phase 3: Plan audits and fixes
- Phase 4: Review and hand off
- Phase 5: Execute Implementation Plan
Steps
Tasks
Subtasks
Actions
Reference: /enhance-markdown Prompt
Phase 5: Execute Implementation Plan
- Entry Check
- Step 5.1 — Load Plan Artifacts
- Step 5.2 — Run context-map
- Step 5.3 — Run convert-plaintext-to-md
- Step 5.4 — Run boost-prompt
- Step 5.5 — Run ai-prompt-engineering-safety-review
- Step 5.6 — Run prompt-builder
- Step 5.7 — Run update-implementation-plan
- Tasks
Phase 6: Verify All Prompts Optimal
- Step 6.1 — Structural Audit
Steps
Tasks" Prompts/<name
Actions
- Step 6.2 — Cross-Reference Validation
Step 6.3 — Conflict Detection
Step 6.4 — Quality Scoring
- Step 6.5 — Write Verification Report
Summary
Per-File Scores
Conflicts Found
Broken References
- Tasks
Phase 7: Cross-Reference Registry
- Step 7.1 — Build Registry
Prompts/*.md (Conversion Targets)
prompts/*.prompt.md (Workflow Prompts)
Conflict Log
- Step 7.2 — Final Verification
- Tasks
Execution Summary
Personas
Personality
Best Practices
Verification Checklist
Dependencies
Subgoals
MCP Servers & Tools
Hooks
Scripts
Related Prompts
Goal
Context
Inputs
Outputs
Rules
Skills Required
Phases
Phase 1: Discover and verify
Phase 2: Plan conversion
Phase 3: Plan audits and fixes
Phase 4: Review and hand off
Phase 5: Execute Implementation Plan
Steps
Tasks
Subtasks
Actions
Reference: /enhance-markdown Prompt
Phase 5: Execute Implementation Plan
Entry Check
Step 5.1 — Load Plan Artifacts
Step 5.2 — Run context-map
Step 5.3 — Run convert-plaintext-to-md
Step 5.4 — Run boost-prompt
Step 5.5 — Run ai-prompt-engineering-safety-review
Step 5.6 — Run prompt-builder
Step 5.7 — Run update-implementation-plan
Tasks
Phase 6: Verify All Prompts Optimal
Step 6.1 — Structural Audit
Steps
Tasks" Prompts/<name
Actions
Step 6.2 — Cross-Reference Validation
Step 6.3 — Conflict Detection
Step 6.4 — Quality Scoring
Step 6.5 — Write Verification Report
Summary
Per-File Scores
Conflicts Found
Broken References
Tasks
Phase 7: Cross-Reference Registry
Step 7.1 — Build Registry
Prompts/*.md (Conversion Targets)
prompts/*.prompt.md (Workflow Prompts)
Conflict Log
Step 7.2 — Final Verification
Tasks
Execution Summary
Personas
Personality
Best Practices
Verification Checklist
Dependencies
Subgoals
MCP Servers & Tools
Hooks
Scripts
Related Prompts

Create a complete plan for converting plaintext prompts to markdown, enhancing prompt quality, and updating related implementation plans.

Create a detailed prompt-library plan> and verification criteria.

Use this prompt when a prompt-library refresh needs planning beforeimplementation. The primary workflow is:1. prompts/context-map2. prompts/convert-plaintext-to-md3. prompts/boost-prompt4. prompts/ai-prompt-engineering-safety-review5. prompts/update-implementation-plan6. prompts/prompt-builder7. .github/prompts/*.mdReference artifacts:- docs/dev-init-comprehensive-plan.md- docs/dev-init-spec.md- docs/prompts-cross-reference-registry.md- docs/prompt-conversion-enhancement-plan.md- projects/Bash/archive/artifacts/context-maps/dev-init.context.json

Inputs

Source inventory for .github/prompts/*.txt
Existing .github/prompts/*.md files
Related .github/prompts/*.prompt.md workflows
Workspace context and implementation constraints

Outputs

A conversion and enhancement plan
A safety and quality review checklist
File references and integration points
An implementation guide with verification steps

Rules

Core rules: templates/rules-core.md

Run context-map before any conversion or planning work.
Keep the pipeline idempotent and safe to re-run.
Preserve the original task intent while improving clarity and structure.
Keep manual safety notes unless they are wrong.
Use git for rollback; do not create backup copies.
Update the implementation plan when scope or requirements change.
Prefer file-backed evidence over inference.

Skills Required

See full table with per-domain purposes: templates/skills-table-core.md| Skill | Purpose | | ------------------------------------- | ---------------------------------------------------------------- | | context-map (prompt) | Map source, destination, and dependency impact before conversion | | convert-plaintext-to-md | Convert plaintext prompts to markdown | | boost-prompt | Improve prompt quality and structure | | ai-prompt-engineering-safety-review | Review prompts for safety and clarity | | update-implementation-plan | Keep the implementation plan current | | prompt-builder | Scaffold new prompt files | | brainstorming | Explore conversion and enhancement approaches | | plans-and-specs | Create detailed specs with file references | | writing-skills | Craft and improve prompts and instructions | | writing-plans | Structure the plan and its sections | | acpx-executor | Execute a prompt via any ACPX provider | | executing-plans | Execute implementation steps in order | | simplify | Keep the output concise and non-redundant |

Phase 1: Discover and verify

Confirm the source files, target prompts, and required templates before planningchanges.

Phase 2: Plan conversion

Define how .github/prompts/*.txt should convert to .github/prompts/*.md and what must bepreserved.

Phase 3: Plan audits and fixes

Specify how to audit boost-prompt, ai-prompt-engineering-safety-review,update-implementation-plan, and prompt-builder with concrete findings.

Phase 4: Review and hand off

Produce a compact completion report with execution steps, risk notes, and exitcriteria.

Phase 5: Execute Implementation Plan

Run the conversion and enhancement pipeline end-to-end. Idempotent — safe to re-run.

Steps

Load context-map, brainstorming, and plans-and-specs.
Verify the conversion and audit templates exist.
Map the prompt-library dependencies.
Plan .github/prompts/*.txt → .github/prompts/*.md conversion.
Plan audits for the related .github/prompts/*.prompt.md files.
Confirm the target prompts reference context-map where relevant.
Produce the final plan and handoff notes.

Tasks

Read the source inventory and target prompt files- [ ] Confirm the conversion and audit templates exist- [ ] Define the TXT→MD conversion scope and idempotency rules- [ ] Capture the feature specs, file references, and safety checks- [ ] List the prompts that need audit and cross-reference updates- [ ] Document the risk items and missing dependency links- [ ] Compile a concise completion report for implementation

Subtasks

Map .github/prompts/*.txt to their .md counterparts- [ ] Preserve source intent and safety notes in each converted file- [ ] Check for context-map references in each audit target- [ ] Record any missing template or format assumptions

Actions

read_file("prompts/context-map.prompt.md") — Load the dependency map before planning
skill_view(name="brainstorming") — Expand workflow options and tradeoffs
skill_view(name="plans-and-specs") — Produce the detailed plan artifact
skill_view(name="acpx-executor") — Dispatch tasks to ACPX providers
search_files(pattern, target) — Locate related prompt files and templates
delegate_task(goal, toolsets) — Split audit work when multiple files are involved---

Reference: `/enhance-markdown` Prompt

Full definition: _archive/dev-init.prompts.txt (lines 99–836)The /enhance-markdown prompt is the four-phase markdown auditor and enhancerused by the TXT→MD conversion pipeline. Key properties:bash/enhance-markdown <file [slug] # audit + enhance mode/enhance-markdown --txt-to-md [file] # TXT→MD conversion mode (batch or single)- 9-section template: Skills → Subagents → Personas → Rules → Phases → Steps → Tasks → Subtasks → Actions- Batch size: exactly 7 files per batch- Resumable: each phase checks for existing artifacts before re-running- Plugin planning: uses createPlan/createSpec/appendSpec with companion markdown fallback---

Phase 5: Execute Implementation Plan

Run the conversion and enhancement pipeline end-to-end. Idempotent — safe to re-run.

Entry Check


IF docs/prompt-conversion-enhancement-plan.md EXISTS → read it, skip to Step 5.3ELSE IF docs/dev-init-comprehensive-plan.md EXISTS → plan exists, skip to Step 5.2ELSE → run Phase 1–4 first, then return here

Step 5.1 — Load Plan Artifacts


bashread_file("docs/dev-init-comprehensive-plan.md")read_file("docs/dev-init-spec.md")read_file("projects/Bash/archive/artifacts/context-maps/dev-init.context.json")

Step 5.2 — Run `context-map`

Load prompts/context-map.prompt.md2. Map all .github/prompts/*.txt → .github/prompts/*.md pairs3. Map all .github/prompts/*.prompt.md dependencies4. Write docs/prompts-cross-reference-registry.md

Step 5.3 — Run `convert-plaintext-to-md`

For each .github/prompts/*.txt file:1. Read raw .txt content2. Apply enhancement (Stanford/Anthropic patterns): - Critical rules in first 15% - Nesting depth ≤ 4 - Instruction ratio 40–50% - Single source of truth (no rule repeated

2×) - 3-tier prioritization: Safety → Core Workflow → Optimization3. Apply 9-section template: Skills → Subagents → Personas → Rules → Phases → Steps → Tasks → Subtasks → Actions4. Write to .github/prompts/<same-stem>.md (overwrite if exists)

Step 5.4 — Run `boost-prompt`

For each .github/prompts/*.md file:1. Load prompts/boost-prompt.prompt.md2. Apply quality enhancements: - Strengthen rule language (imperative, specific) - Add missing frontmatter (title, description, tags, trigger) - Ensure consistent heading hierarchy - Remove redundant sections3. Constraint Preservation Audit: log any rule removals with justification4. Write enhanced file back

Step 5.5 — Run `ai-prompt-engineering-safety-review`

For each .github/prompts/*.md and .github/prompts/*.prompt.md:1. Load prompts/ai-prompt-engineering-safety-review.prompt.md2. Check for: - Credential handling safety - Backup/rollback instructions present - Approval workflows for destructive ops - No fabricated verification claims3. Add safety notes where missing4. Fail if critical safety constraints removed — halt and report

Step 5.6 — Run `prompt-builder`

For any missing prompts:1. Load prompts/prompt-builder.prompt.md2. Scaffold new .prompt.md files from template3. Populate with content from plan artifacts

Step 5.7 — Run `update-implementation-plan`

Load prompts/update-implementation-plan.prompt.md2. Update docs/dev-init-comprehensive-plan.md with: - Actual files processed - Issues found and resolved - Remaining items3. Mark completed phases

Tasks

Load all plan artifacts (Step 5.1)- [ ] Run context-map and write cross-reference registry (Step 5.2)- [ ] Convert all .github/prompts/*.txt → .github/prompts/*.md (Step 5.3)- [ ] Boost all .github/prompts/*.md files (Step 5.4)- [ ] Safety-review all prompts (Step 5.5)- [ ] Build missing prompts (Step 5.6)- [ ] Update implementation plan (Step 5.7)---

Phase 6: Verify All Prompts Optimal

Independent verification — read Phase 1 outputs only, not Phase 5 outputs (prevents confirmation bias).

Step 6.1 — Structural Audit

For each .github/prompts/*.md:


## Check frontmatterhead -10 Prompts/<name

> .prompts.md | grep "^title:"head -10 Prompts/<name>.prompts.md | grep "^description:"head -10 Prompts/<name>.prompts.md | grep "^trigger:"head -10 Prompts/<name>.prompts.md | grep "^tags:"# Check required sections existgrep -c "^

## Steps

" Prompts/<name>.prompts.mdgrep -c "^

## Tasks" Prompts/<name

> .prompts.mdgrep -c "^

## Actions

" Prompts/<name>.prompts.md

Step 6.2 — Cross-Reference Validation


## Verify all internal references resolvegrep -o '\[.*\](\./[^)]*)' Prompts/<name>.prompts.md | while read ref; do path=$(echo "$ref" | sed 's/.*(\(.*\))/\1/') test -f "$path" || echo "BROKEN REF: $ref in Prompts/<name>.prompts.md"done

Step 6.3 — Conflict Detection


## Check for duplicate triggers across Prompts/grep -h "^trigger:" Prompts/*.prompts.md | sort | uniq -d# Check for duplicate titlesgrep -h "^title:" Prompts/*.prompts.md | sort | uniq -d

Step 6.4 — Quality Scoring

Score each .github/prompts/*.md:| Criterion | Points | Detection || ------------------------------------

| ------ | ------------------------------------------------ || YAML frontmatter complete | +20 | title + description + trigger + tags present || Summary paragraph present | +15 | First paragraph after frontmatter is a summary || All required sections | +20 | Goal, Phases, Steps, Tasks, Actions present || Skills section references real skills | +10 | Each skill listed exists in hermes skills list || Actions use real tools | +10 | Each action maps to an available tool || No broken internal refs | +15 | All relative paths resolve || Consistent heading hierarchy | +10 | H1→H2→H3, no skipped levels |Threshold: ≥80 = optimal, 60–79 = needs work, <60 = rewrite required.

Step 6.5 — Write Verification Report


## docs/dev-init-verification-report.md

## Summary

- Total .md files: N- Optimal (≥80): N- Needs work (60–79): N- Rewrite required (<60): N

## Per-File Scores

| File | Score | Issues || ---- | ----- | ------ || ... | ... | ... |

## Conflicts Found

| Type | Files | Resolution || ----------------

- | ----- | ---------- || Duplicate trigger | ... | ... |

## Broken References

| File | Reference | Status || ---

- | --------- | ------ || ... | ... | ... |

Tasks

Structural audit all .github/prompts/*.md files- [ ] Validate all cross-references resolve- [ ] Detect duplicate triggers/titles- [ ] Score each file against quality criteria- [ ] Write verification report to docs/dev-init-verification-report.md- [ ] Fix any files scoring <80- [ ] Re-verify after fixes---

Phase 7: Cross-Reference Registry

Build the master registry of all prompts and their relationships.

Step 7.1 — Build Registry

Write docs/prompts-cross-reference-registry.md:```markdown# Prompts Cross-Reference Registry

Prompts/*.md (Conversion Targets)

| File | Trigger | Status | Score | Depends On || --------------------

| ----------- | ------ | ----- | ------------------------------------ || skills-fix.prompts.md | /skills-fix | ✅ | 85 | context-map, skill-judge || dev-init.prompts.md | /dev-init | ✅ | 90 | context-map, convert-plaintext-to-md || ... | ... | ... | ... | ... |

prompts/*.prompt.md (Workflow Prompts)

| File | Purpose | Used By || --------------------------------

| ------------------- | ------------------- || context-map.prompt.md | Dependency mapping | All dev-init phases || convert-plaintext-to-md.prompt.md | TXT→MD conversion | Phase 5.3 || boost-prompt.prompt.md | Quality enhancement | Phase 5.4 || ... | ... | ... |

Conflict Log

| Conflict | Files | Resolution || -------

| ----- | ---------- || ... | ... | ... |```

Step 7.2 — Final Verification


bash# Count total promptsls Prompts/*.prompts.md | wc -lls prompts/*.prompt.md | wc -l# Verify registry matches disk# Every file on disk is in registry# Every registry entry exists on disk

Tasks

Build cross-reference registry- [ ] Verify registry matches disk state- [ ] Commit all changes with git commit -m "dev-init: complete prompt-library refresh"---

Execution Summary

When all 7 phases are complete, output:========================================DEV INIT — PROMPT LIBRARY REFRESH COMPLETE========================================Phases executed: 7/7Prompts converted: N/N (.txt → .md)Prompts boosted: N/NSafety reviews: N passed, N failedQuality score ≥80: N/NConflicts resolved: NBroken refs fixed: NRegistry: docs/prompts-cross-reference-registry.mdVerification: docs/dev-init-verification-report.md========================================

Personas

See templates/personas.md for shared persona templates.

Persona	When to Use
Developer	Implementation, debugging, refactoring
Reviewer	Code review, quality assurance
User	General purpose, operations

Personality

See templates/personality.md for shared personality guidelines.

Tone: Direct, practical, actionable
Style: Structured with clear steps and verification
Avoid: Ambiguity, assumptions, scope creep
Encourage: Evidence-based decisions, minimal changes

Best Practices

See templates/best-practices.md for cross-cutting best practices.

DRY — Reference shared templates instead of duplicating content.
Structured output — Use clear sections with consistent heading levels.
Verification gates — Always verify before claiming completion.
Minimal changes — Fix root cause, not symptoms.

Verification Checklist

#	Gate	Criterion
1	Scope	Change matches the original request
2	Quality	Meets project standards
3	Tests	Tests pass (if applicable)
4	Regression	No unintended side effects
5	Docs	Changes documented if needed

Dependencies

See templates/deps-core.md for shared dependency patterns.

Subgoals

Prepare — Understand requirements and prerequisites.
Execute — Follow structured workflow with incremental progress.
Verify — Confirm output meets requirements and standards.
Document — Record results, decisions, and lessons learned.

MCP Servers & Tools

The following MCP servers and tools are available for this task. Use them in preference to native equivalents per MCP-first tooling policy.

Hooks

Shared workspace hooks run around this prompt's execution — see .github/hooks/README.md: session-logger, session-auto-commit, governance-audit, pre-exec-validate.sh, post-exec-state-log.py.

Scripts

Prompt-library tooling (see .enhance/):

.enhance/analyze_prompts.py — prompt-library analyzer (Phase 5/7 gate)
.enhance/verify_phase3.py, .enhance/fix_class_e.py, .enhance/fix_frontmatter_plan.py — Class C–E repair/verify tooling
.github/hooks/* — hook implementations referenced in the Hooks section

Related Prompts

Workflow

Same-family prompts:

--- name: dev-init title: Development Initialization description: Initialize a project for prompt-driven development: discover, plan conversion, audit/fix prompts, execute an implementation plan, and verify. trigger: /dev-init category: development version: 1.0.0 author: Hermes Agent tags: metadata: hermes: profile: code-architect priority: medium copilot: model_required: sonnet opencode: enabled: true codex: toolsets: skills: - skill: using-superpowers dependencies: [] formatter: markdown license: MIT --- ## Table of Contents ## Goal Initialize a project for prompt-driven development: discover, plan conversion, audit/fix prompts, execute an implementation plan, and verify. ## Context ## Phases # Table of Contents - [Goal](#goal) - [Context](#context) - [Inputs](#inputs) - [Outputs](#outputs) - [Rules](#rules) - [Skills Required](#skills-required) - [Phases](#phases) - [Phase 1: Discover and verify](#phase-1:-discover-and-verify) - [Phase 2: Plan conversion](#phase-2:-plan-conversion) - [Phase 3: Plan audits and fixes](#phase-3:-plan-audits-and-fixes) - [Phase 4: Review and hand off](#phase-4:-review-and-hand-off) - [Phase 5: Execute Implementation Plan](#phase-5:-execute-implementation-plan) - [Steps](#steps) - [Tasks](#tasks) - [Subtasks](#subtasks) - [Actions](#actions) - [Reference: `/enhance-markdown` Prompt](#reference:-`/enhance-markdown`-prompt) - [Phase 5: Execute Implementation Plan](#phase-5:-execute-implementation-plan) - [Entry Check](#entry-check) - [Step 5.1 — Load Plan Artifacts](#step-51-—-load-plan-artifacts) - [Step 5.2 — Run `context-map`](#step-52-—-run-`context-map`) - [Step 5.3 — Run `convert-plaintext-to-md`](#step-53-—-run-`convert-plaintext-to-md`) - [Step 5.4 — Run `boost-prompt`](#step-54-—-run-`boost-prompt`) - [Step 5.5 — Run `ai-prompt-engineering-safety-review`](#step-55-—-run-`ai-prompt-engineering-safety-review`) - [Step 5.6 — Run `prompt-builder`](#step-56-—-run-`prompt-builder`) - [Step 5.7 — Run `update-implementation-plan`](#step-57-—-run-`update-implementation-plan`) - [Tasks](#tasks) - [Phase 6: Verify All Prompts Optimal](#phase-6:-verify-all-prompts-optimal) - [Step 6.1 — Structural Audit](#step-61-—-structural-audit) - [Steps](#steps) - [Tasks" Prompts/<name](#tasks"-prompts/<name) - [Actions](#actions) - [Step 6.2 — Cross-Reference Validation](#step-62-—-cross-reference-validation) - [Step 6.3 — Conflict Detection](#step-63-—-conflict-detection) - [Step 6.4 — Quality Scoring](#step-64-—-quality-scoring) - [Step 6.5 — Write Verification Report](#step-65-—-write-verification-report) - [Summary](#summary) - [Per-File Scores](#per-file-scores) - [Conflicts Found](#conflicts-found) - [Broken References](#broken-references) - [Tasks](#tasks) - [Phase 7: Cross-Reference Registry](#phase-7:-cross-reference-registry) - [Step 7.1 — Build Registry](#step-71-—-build-registry) - [Prompts/\*.md (Conversion Targets)](#prompts/\*md-conversion-targets) - [prompts/\*.prompt.md (Workflow Prompts)](#prompts/\*promptmd-workflow-prompts) - [Conflict Log](#conflict-log) - [Step 7.2 — Final Verification](#step-72-—-final-verification) - [Tasks](#tasks) - [Execution Summary](#execution-summary) - [Personas](#personas) - [Personality](#personality) - [Best Practices](#best-practices) - [Verification Checklist](#verification-checklist) - [Dependencies](#dependencies) - [Subgoals](#subgoals) - [MCP Servers & Tools](#mcp-servers-&-tools) - [Hooks](#hooks) - [Scripts](#scripts) - [Related Prompts](#related-prompts) - [Goal](#goal) - [Context](#context) - [Inputs](#inputs) - [Outputs](#outputs) - [Rules](#rules) - [Skills Required](#skills-required) - [Phases](#phases) - [Phase 1: Discover and verify](#phase-1:-discover-and-verify) - [Phase 2: Plan conversion](#phase-2:-plan-conversion) - [Phase 3: Plan audits and fixes](#phase-3:-plan-audits-and-fixes) - [Phase 4: Review and hand off](#phase-4:-review-and-hand-off) - [Phase 5: Execute Implementation Plan](#phase-5:-execute-implementation-plan) - [Steps](#steps) - [Tasks](#tasks) - [Subtasks](#subtasks) - [Actions](#actions) - [Reference: `/enhance-markdown` Prompt](#reference:-`/enhance-markdown`-prompt) - [Phase 5: Execute Implementation Plan](#phase-5:-execute-implementation-plan) - [Entry Check](#entry-check) - [Step 5.1 — Load Plan Artifacts](#step-51-—-load-plan-artifacts) - [Step 5.2 — Run `context-map`](#step-52-—-run-`context-map`) - [Step 5.3 — Run `convert-plaintext-to-md`](#step-53-—-run-`convert-plaintext-to-md`) - [Step 5.4 — Run `boost-prompt`](#step-54-—-run-`boost-prompt`) - [Step 5.5 — Run `ai-prompt-engineering-safety-review`](#step-55-—-run-`ai-prompt-engineering-safety-review`) - [Step 5.6 — Run `prompt-builder`](#step-56-—-run-`prompt-builder`) - [Step 5.7 — Run `update-implementation-plan`](#step-57-—-run-`update-implementation-plan`) - [Tasks](#tasks) - [Phase 6: Verify All Prompts Optimal](#phase-6:-verify-all-prompts-optimal) - [Step 6.1 — Structural Audit](#step-61-—-structural-audit) - [Steps](#steps) - [Tasks" Prompts/<name](#tasks"-prompts/<name) - [Actions](#actions) - [Step 6.2 — Cross-Reference Validation](#step-62-—-cross-reference-validation) - [Step 6.3 — Conflict Detection](#step-63-—-conflict-detection) - [Step 6.4 — Quality Scoring](#step-64-—-quality-scoring) - [Step 6.5 — Write Verification Report](#step-65-—-write-verification-report) - [Summary](#summary) - [Per-File Scores](#per-file-scores) - [Conflicts Found](#conflicts-found) - [Broken References](#broken-references) - [Tasks](#tasks) - [Phase 7: Cross-Reference Registry](#phase-7:-cross-reference-registry) - [Step 7.1 — Build Registry](#step-71-—-build-registry) - [Prompts/\*.md (Conversion Targets)](#prompts/\*md-conversion-targets) - [prompts/\*.prompt.md (Workflow Prompts)](#prompts/\*promptmd-workflow-prompts) - [Conflict Log](#conflict-log) - [Step 7.2 — Final Verification](#step-72-—-final-verification) - [Tasks](#tasks) - [Execution Summary](#execution-summary) - [Personas](#personas) - [Personality](#personality) - [Best Practices](#best-practices) - [Verification Checklist](#verification-checklist) - [Dependencies](#dependencies) - [Subgoals](#subgoals) - [MCP Servers & Tools](#mcp-servers-&-tools) - [Hooks](#hooks) - [Scripts](#scripts) - [Related Prompts](#related-prompts) Create a complete plan for converting plaintext prompts to markdown, enhancing prompt quality, and updating related implementation plans. >> Create a detailed prompt-library plan> and verification criteria. Use this prompt when a prompt-library refresh needs planning beforeimplementation. The primary workflow is:1. `prompts/context-map`2. `prompts/convert-plaintext-to-md`3. `prompts/boost-prompt`4. `prompts/ai-prompt-engineering-safety-review`5. `prompts/update-implementation-plan`6. `prompts/prompt-builder`7. `.github/prompts/*.md`Reference artifacts:- `docs/dev-init-comprehensive-plan.md`- `docs/dev-init-spec.md`- `docs/prompts-cross-reference-registry.md`- `docs/prompt-conversion-enhancement-plan.md`- `projects/Bash/archive/artifacts/context-maps/dev-init.context.json` ## Inputs - Source inventory for `.github/prompts/*.txt` - Existing `.github/prompts/*.md` files - Related `.github/prompts/*.prompt.md` workflows - Workspace context and implementation constraints ## Outputs - A conversion and enhancement plan - A safety and quality review checklist - File references and integration points - An implementation guide with verification steps ## Rules > Core rules: [`templates/rules-core.md`](templates/rules-core.md) 1. Run `context-map` before any conversion or planning work. 2. Keep the pipeline idempotent and safe to re-run. 3. Preserve the original task intent while improving clarity and structure. 4. Keep manual safety notes unless they are wrong. 5. Use git for rollback; do not create backup copies. 6. Update the implementation plan when scope or requirements change. 7. Prefer file-backed evidence over inference. ## Skills Required > See full table with per-domain purposes: > [`templates/skills-table-core.md`](templates/skills-table-core.md)| Skill | Purpose | | ------------------------------------- | ---------------------------------------------------------------- | | `context-map` (prompt) | Map source, destination, and dependency impact before conversion | | `convert-plaintext-to-md` | Convert plaintext prompts to markdown | | `boost-prompt` | Improve prompt quality and structure | | `ai-prompt-engineering-safety-review` | Review prompts for safety and clarity | | `update-implementation-plan` | Keep the implementation plan current | | `prompt-builder` | Scaffold new prompt files | | `brainstorming` | Explore conversion and enhancement approaches | | `plans-and-specs` | Create detailed specs with file references | | `writing-skills` | Craft and improve prompts and instructions | | `writing-plans` | Structure the plan and its sections | | `acpx-executor` | Execute a prompt via any ACPX provider | | `executing-plans` | Execute implementation steps in order | | `simplify` | Keep the output concise and non-redundant | ### Phase 1: Discover and verify Confirm the source files, target prompts, and required templates before planningchanges. ### Phase 2: Plan conversion Define how `.github/prompts/*.txt` should convert to `.github/prompts/*.md` and what must bepreserved. ### Phase 3: Plan audits and fixes Specify how to audit `boost-prompt`, `ai-prompt-engineering-safety-review`,`update-implementation-plan`, and `prompt-builder` with concrete findings. ### Phase 4: Review and hand off Produce a compact completion report with execution steps, risk notes, and exitcriteria. ### Phase 5: Execute Implementation Plan Run the conversion and enhancement pipeline end-to-end. Idempotent — safe to re-run. ## Steps 1. Load `context-map`, `brainstorming`, and `plans-and-specs`. 2. Verify the conversion and audit templates exist. 3. Map the prompt-library dependencies. 4. Plan `.github/prompts/*.txt` → `.github/prompts/*.md` conversion. 5. Plan audits for the related `.github/prompts/*.prompt.md` files. 6. Confirm the target prompts reference `context-map` where relevant. 7. Produce the final plan and handoff notes. ## Tasks - [ ] Read the source inventory and target prompt files- [ ] Confirm the conversion and audit templates exist- [ ] Define the TXT→MD conversion scope and idempotency rules- [ ] Capture the feature specs, file references, and safety checks- [ ] List the prompts that need audit and cross-reference updates- [ ] Document the risk items and missing dependency links- [ ] Compile a concise completion report for implementation ## Subtasks - [ ] Map `.github/prompts/*.txt` to their `.md` counterparts- [ ] Preserve source intent and safety notes in each converted file- [ ] Check for `context-map` references in each audit target- [ ] Record any missing template or format assumptions ## Actions - `read_file("prompts/context-map.prompt.md")` — Load the dependency map before planning - `skill_view(name="brainstorming")` — Expand workflow options and tradeoffs - `skill_view(name="plans-and-specs")` — Produce the detailed plan artifact - `skill_view(name="acpx-executor")` — Dispatch tasks to ACPX providers - `search_files(pattern, target)` — Locate related prompt files and templates - `delegate_task(goal, toolsets)` — Split audit work when multiple files are involved--- ## Reference: `/enhance-markdown` Prompt > Full definition: `_archive/dev-init.prompts.txt` (lines 99–836)The `/enhance-markdown` prompt is the four-phase markdown auditor and enhancerused by the TXT→MD conversion pipeline. Key properties:```bash/enhance-markdown <file > [slug] # audit + enhance mode/enhance-markdown --txt-to-md [file] # TXT→MD conversion mode (batch or single)```- **9-section template**: Skills → Subagents → Personas → Rules → Phases → Steps → Tasks → Subtasks → Actions- **Batch size**: exactly 7 files per batch- **Resumable**: each phase checks for existing artifacts before re-running- **Plugin planning**: uses `createPlan`/`createSpec`/`appendSpec` with companion markdown fallback--- ## Phase 5: Execute Implementation Plan > Run the conversion and enhancement pipeline end-to-end. Idempotent — safe to re-run. ### Entry Check ``` IF docs/prompt-conversion-enhancement-plan.md EXISTS → read it, skip to Step 5.3ELSE IF docs/dev-init-comprehensive-plan.md EXISTS → plan exists, skip to Step 5.2ELSE → run Phase 1–4 first, then return here ``` ### Step 5.1 — Load Plan Artifacts ``` bashread_file("docs/dev-init-comprehensive-plan.md")read_file("docs/dev-init-spec.md")read_file("projects/Bash/archive/artifacts/context-maps/dev-init.context.json") ``` ### Step 5.2 — Run `context-map` 1. Load `prompts/context-map.prompt.md`2. Map all `.github/prompts/*.txt` → `.github/prompts/*.md` pairs3. Map all `.github/prompts/*.prompt.md` dependencies4. Write `docs/prompts-cross-reference-registry.md` ### Step 5.3 — Run `convert-plaintext-to-md` For each `.github/prompts/*.txt` file:1. Read raw `.txt` content2. Apply enhancement (Stanford/Anthropic patterns): - Critical rules in first 15% - Nesting depth ≤ 4 - Instruction ratio 40–50% - Single source of truth (no rule repeated > 2×) - 3-tier prioritization: Safety → Core Workflow → Optimization3. Apply 9-section template: `Skills → Subagents → Personas → Rules → Phases → Steps → Tasks → Subtasks → Actions`4. Write to `.github/prompts/<same-stem>.md` (overwrite if exists) ### Step 5.4 — Run `boost-prompt` For each `.github/prompts/*.md` file:1. Load `prompts/boost-prompt.prompt.md`2. Apply quality enhancements: - Strengthen rule language (imperative, specific) - Add missing frontmatter (title, description, tags, trigger) - Ensure consistent heading hierarchy - Remove redundant sections3. **Constraint Preservation Audit**: log any rule removals with justification4. Write enhanced file back ### Step 5.5 — Run `ai-prompt-engineering-safety-review` For each `.github/prompts/*.md` and `.github/prompts/*.prompt.md`:1. Load `prompts/ai-prompt-engineering-safety-review.prompt.md`2. Check for: - Credential handling safety - Backup/rollback instructions present - Approval workflows for destructive ops - No fabricated verification claims3. Add safety notes where missing4. **Fail if critical safety constraints removed** — halt and report ### Step 5.6 — Run `prompt-builder` For any missing prompts:1. Load `prompts/prompt-builder.prompt.md`2. Scaffold new `.prompt.md` files from template3. Populate with content from plan artifacts ### Step 5.7 — Run `update-implementation-plan` 1. Load `prompts/update-implementation-plan.prompt.md`2. Update `docs/dev-init-comprehensive-plan.md` with: - Actual files processed - Issues found and resolved - Remaining items3. Mark completed phases ### Tasks - [ ] Load all plan artifacts (Step 5.1)- [ ] Run context-map and write cross-reference registry (Step 5.2)- [ ] Convert all `.github/prompts/*.txt` → `.github/prompts/*.md` (Step 5.3)- [ ] Boost all `.github/prompts/*.md` files (Step 5.4)- [ ] Safety-review all prompts (Step 5.5)- [ ] Build missing prompts (Step 5.6)- [ ] Update implementation plan (Step 5.7)--- ## Phase 6: Verify All Prompts Optimal > Independent verification — read Phase 1 outputs only, not Phase 5 outputs (prevents confirmation bias). ### Step 6.1 — Structural Audit For each `.github/prompts/*.md`: ```bash ## Check frontmatterhead -10 Prompts/<name > .prompts.md | grep "^title:"head -10 Prompts/<name>.prompts.md | grep "^description:"head -10 Prompts/<name>.prompts.md | grep "^trigger:"head -10 Prompts/<name>.prompts.md | grep "^tags:"# Check required sections existgrep -c "^ ## Steps " Prompts/<name>.prompts.mdgrep -c "^ ## Tasks" Prompts/<name > .prompts.mdgrep -c "^ ## Actions " Prompts/<name>.prompts.md ``` ### Step 6.2 — Cross-Reference Validation ```bash ## Verify all internal references resolvegrep -o '\[.*\](\./[^)]*)' Prompts/<name>.prompts.md | while read ref; do path=$(echo "$ref" | sed 's/.*($.*$)/\1/') test -f "$path" || echo "BROKEN REF: $ref in Prompts/<name>.prompts.md"done ``` ## Step 6.3 — Conflict Detection ```bash ## Check for duplicate triggers across Prompts/grep -h "^trigger:" Prompts/*.prompts.md | sort | uniq -d# Check for duplicate titlesgrep -h "^title:" Prompts/*.prompts.md | sort | uniq -d ``` ## Step 6.4 — Quality Scoring Score each `.github/prompts/*.md`:| Criterion | Points | Detection || ------------------------------------ - | ------ | ------------------------------------------------ || YAML frontmatter complete | +20 | title + description + trigger + tags present || Summary paragraph present | +15 | First paragraph after frontmatter is a summary || All required sections | +20 | Goal, Phases, Steps, Tasks, Actions present || Skills section references real skills | +10 | Each skill listed exists in `hermes skills list` || Actions use real tools | +10 | Each action maps to an available tool || No broken internal refs | +15 | All relative paths resolve || Consistent heading hierarchy | +10 | H1→H2→H3, no skipped levels |**Threshold**: ≥80 = optimal, 60–79 = needs work, <60 = rewrite required. ### Step 6.5 — Write Verification Report ```markdown ## docs/dev-init-verification-report.md ## Summary - Total .md files: N- Optimal (≥80): N- Needs work (60–79): N- Rewrite required (<60): N ## Per-File Scores | File | Score | Issues || ---- | ----- | ------ || ... | ... | ... | ## Conflicts Found | Type | Files | Resolution || ---------------- - | ----- | ---------- || Duplicate trigger | ... | ... | ## Broken References | File | Reference | Status || --- - | --------- | ------ || ... | ... | ... | ``` ### Tasks - [ ] Structural audit all `.github/prompts/*.md` files- [ ] Validate all cross-references resolve- [ ] Detect duplicate triggers/titles- [ ] Score each file against quality criteria- [ ] Write verification report to `docs/dev-init-verification-report.md`- [ ] Fix any files scoring <80- [ ] Re-verify after fixes--- ## Phase 7: Cross-Reference Registry > Build the master registry of all prompts and their relationships. ### Step 7.1 — Build Registry Write `docs/prompts-cross-reference-registry.md`:```markdown# Prompts Cross-Reference Registry ## Prompts/\*.md (Conversion Targets) | File | Trigger | Status | Score | Depends On || -------------------- - | ----------- | ------ | ----- | ------------------------------------ || skills-fix.prompts.md | /skills-fix | ✅ | 85 | context-map, skill-judge || dev-init.prompts.md | /dev-init | ✅ | 90 | context-map, convert-plaintext-to-md || ... | ... | ... | ... | ... | ## prompts/\*.prompt.md (Workflow Prompts) | File | Purpose | Used By || -------------------------------- - | ------------------- | ------------------- || context-map.prompt.md | Dependency mapping | All dev-init phases || convert-plaintext-to-md.prompt.md | TXT→MD conversion | Phase 5.3 || boost-prompt.prompt.md | Quality enhancement | Phase 5.4 || ... | ... | ... | ## Conflict Log | Conflict | Files | Resolution || ------- - | ----- | ---------- || ... | ... | ... |``` ### Step 7.2 — Final Verification ``` bash# Count total promptsls Prompts/*.prompts.md | wc -lls prompts/*.prompt.md | wc -l# Verify registry matches disk# Every file on disk is in registry# Every registry entry exists on disk ``` ### Tasks - [ ] Build cross-reference registry- [ ] Verify registry matches disk state- [ ] Commit all changes with `git commit -m "dev-init: complete prompt-library refresh"`--- ## Execution Summary When all 7 phases are complete, output:```========================================DEV INIT — PROMPT LIBRARY REFRESH COMPLETE========================================Phases executed: 7/7Prompts converted: N/N (.txt → .md)Prompts boosted: N/NSafety reviews: N passed, N failedQuality score ≥80: N/NConflicts resolved: NBroken refs fixed: NRegistry: docs/prompts-cross-reference-registry.mdVerification: docs/dev-init-verification-report.md===

{ "prompt": "---\nname: dev-init\ntitle: Development Initialization\ndescription: Initialize a project for prompt-driven development: discover, plan conversion, audit/fix prompts, execute an implementation plan, and verify.\ntrigger: /dev-init\ncategory: development\nversion: 1.0.0\nauthor: Hermes Agent\ntags: \nmetadata: \nhermes: \nprofile: code-architect\npriority: medium\ncopilot: \nmodel_required: sonnet\nopencode: \nenabled: true\ncodex: \ntoolsets: \nskills: \n- skill: using-superpowers\ndependencies: []\nformatter: markdown\nlicense: MIT\n---\n\n## Table of Contents\n\n## Goal\nInitialize a project for prompt-driven development: discover, plan conversion, audit/fix prompts, execute an implementation plan, and verify.\n\n## Context\n\n## Phases\n\n\n\n# Table of Contents\n\n- [Goal](#goal)\n- [Context](#context)\n- [Inputs](#inputs)\n- [Outputs](#outputs)\n- [Rules](#rules)\n- [Skills Required](#skills-required)\n- [Phases](#phases)\n - [Phase 1: Discover and verify](#phase-1:-discover-and-verify)\n - [Phase 2: Plan conversion](#phase-2:-plan-conversion)\n - [Phase 3: Plan audits and fixes](#phase-3:-plan-audits-and-fixes)\n - [Phase 4: Review and hand off](#phase-4:-review-and-hand-off)\n - [Phase 5: Execute Implementation Plan](#phase-5:-execute-implementation-plan)\n- [Steps](#steps)\n- [Tasks](#tasks)\n- [Subtasks](#subtasks)\n- [Actions](#actions)\n- [Reference: `/enhance-markdown` Prompt](#reference:-`/enhance-markdown`-prompt)\n- [Phase 5: Execute Implementation Plan](#phase-5:-execute-implementation-plan)\n - [Entry Check](#entry-check)\n - [Step 5.1 — Load Plan Artifacts](#step-51-—-load-plan-artifacts)\n - [Step 5.2 — Run `context-map`](#step-52-—-run-`context-map`)\n - [Step 5.3 — Run `convert-plaintext-to-md`](#step-53-—-run-`convert-plaintext-to-md`)\n - [Step 5.4 — Run `boost-prompt`](#step-54-—-run-`boost-prompt`)\n - [Step 5.5 — Run `ai-prompt-engineering-safety-review`](#step-55-—-run-`ai-prompt-engineering-safety-review`)\n - [Step 5.6 — Run `prompt-builder`](#step-56-—-run-`prompt-builder`)\n - [Step 5.7 — Run `update-implementation-plan`](#step-57-—-run-`update-implementation-plan`)\n - [Tasks](#tasks)\n- [Phase 6: Verify All Prompts Optimal](#phase-6:-verify-all-prompts-optimal)\n - [Step 6.1 — Structural Audit](#step-61-—-structural-audit)\n- [Steps](#steps)\n- [Tasks\" Prompts/<name](#tasks\"-prompts/<name)\n- [Actions](#actions)\n - [Step 6.2 — Cross-Reference Validation](#step-62-—-cross-reference-validation)\n- [Step 6.3 — Conflict Detection](#step-63-—-conflict-detection)\n- [Step 6.4 — Quality Scoring](#step-64-—-quality-scoring)\n - [Step 6.5 — Write Verification Report](#step-65-—-write-verification-report)\n- [Summary](#summary)\n- [Per-File Scores](#per-file-scores)\n- [Conflicts Found](#conflicts-found)\n- [Broken References](#broken-references)\n - [Tasks](#tasks)\n- [Phase 7: Cross-Reference Registry](#phase-7:-cross-reference-registry)\n - [Step 7.1 — Build Registry](#step-71-—-build-registry)\n- [Prompts/\\*.md (Conversion Targets)](#prompts/\\*md-conversion-targets)\n- [prompts/\\*.prompt.md (Workflow Prompts)](#prompts/\\*promptmd-workflow-prompts)\n- [Conflict Log](#conflict-log)\n - [Step 7.2 — Final Verification](#step-72-—-final-verification)\n - [Tasks](#tasks)\n- [Execution Summary](#execution-summary)\n- [Personas](#personas)\n- [Personality](#personality)\n- [Best Practices](#best-practices)\n- [Verification Checklist](#verification-checklist)\n- [Dependencies](#dependencies)\n- [Subgoals](#subgoals)\n- [MCP Servers & Tools](#mcp-servers-&-tools)\n- [Hooks](#hooks)\n- [Scripts](#scripts)\n- [Related Prompts](#related-prompts)\n\n\n\n- [Goal](#goal)\n- [Context](#context)\n- [Inputs](#inputs)\n- [Outputs](#outputs)\n- [Rules](#rules)\n- [Skills Required](#skills-required)\n- [Phases](#phases)\n- [Phase 1: Discover and verify](#phase-1:-discover-and-verify)\n- [Phase 2: Plan conversion](#phase-2:-plan-conversion)\n- [Phase 3: Plan audits and fixes](#phase-3:-plan-audits-and-fixes)\n- [Phase 4: Review and hand off](#phase-4:-review-and-hand-off)\n- [Phase 5: Execute Implementation Plan](#phase-5:-execute-implementation-plan)\n- [Steps](#steps)\n- [Tasks](#tasks)\n- [Subtasks](#subtasks)\n- [Actions](#actions)\n- [Reference: `/enhance-markdown` Prompt](#reference:-`/enhance-markdown`-prompt)\n- [Phase 5: Execute Implementation Plan](#phase-5:-execute-implementation-plan)\n- [Entry Check](#entry-check)\n- [Step 5.1 — Load Plan Artifacts](#step-51-—-load-plan-artifacts)\n- [Step 5.2 — Run `context-map`](#step-52-—-run-`context-map`)\n- [Step 5.3 — Run `convert-plaintext-to-md`](#step-53-—-run-`convert-plaintext-to-md`)\n- [Step 5.4 — Run `boost-prompt`](#step-54-—-run-`boost-prompt`)\n- [Step 5.5 — Run `ai-prompt-engineering-safety-review`](#step-55-—-run-`ai-prompt-engineering-safety-review`)\n- [Step 5.6 — Run `prompt-builder`](#step-56-—-run-`prompt-builder`)\n- [Step 5.7 — Run `update-implementation-plan`](#step-57-—-run-`update-implementation-plan`)\n- [Tasks](#tasks)\n- [Phase 6: Verify All Prompts Optimal](#phase-6:-verify-all-prompts-optimal)\n- [Step 6.1 — Structural Audit](#step-61-—-structural-audit)\n- [Steps](#steps)\n- [Tasks\" Prompts/<name](#tasks\"-prompts/<name)\n- [Actions](#actions)\n- [Step 6.2 — Cross-Reference Validation](#step-62-—-cross-reference-validation)\n- [Step 6.3 — Conflict Detection](#step-63-—-conflict-detection)\n- [Step 6.4 — Quality Scoring](#step-64-—-quality-scoring)\n- [Step 6.5 — Write Verification Report](#step-65-—-write-verification-report)\n- [Summary](#summary)\n- [Per-File Scores](#per-file-scores)\n- [Conflicts Found](#conflicts-found)\n- [Broken References](#broken-references)\n- [Tasks](#tasks)\n- [Phase 7: Cross-Reference Registry](#phase-7:-cross-reference-registry)\n- [Step 7.1 — Build Registry](#step-71-—-build-registry)\n- [Prompts/\\*.md (Conversion Targets)](#prompts/\\*md-conversion-targets)\n- [prompts/\\*.prompt.md (Workflow Prompts)](#prompts/\\*promptmd-workflow-prompts)\n- [Conflict Log](#conflict-log)\n- [Step 7.2 — Final Verification](#step-72-—-final-verification)\n- [Tasks](#tasks)\n- [Execution Summary](#execution-summary)\n- [Personas](#personas)\n- [Personality](#personality)\n- [Best Practices](#best-practices)\n- [Verification Checklist](#verification-checklist)\n- [Dependencies](#dependencies)\n- [Subgoals](#subgoals)\n- [MCP Servers & Tools](#mcp-servers-&-tools)\n- [Hooks](#hooks)\n- [Scripts](#scripts)\n- [Related Prompts](#related-prompts)\n\n\n\n\n\nCreate a complete plan for converting plaintext prompts to markdown, enhancing prompt quality, and updating related implementation plans.\n\n>> Create a detailed prompt-library plan> and verification criteria.\n\n\nUse this prompt when a prompt-library refresh needs planning beforeimplementation. The primary workflow is:1. `prompts/context-map`2. `prompts/convert-plaintext-to-md`3. `prompts/boost-prompt`4. `prompts/ai-prompt-engineering-safety-review`5. `prompts/update-implementation-plan`6. `prompts/prompt-builder`7. `.github/prompts/*.md`Reference artifacts:- `docs/dev-init-comprehensive-plan.md`- `docs/dev-init-spec.md`- `docs/prompts-cross-reference-registry.md`- `docs/prompt-conversion-enhancement-plan.md`- `projects/Bash/archive/artifacts/context-maps/dev-init.context.json`\n\n## Inputs\n\n- Source inventory for `.github/prompts/*.txt`\n- Existing `.github/prompts/*.md` files\n- Related `.github/prompts/*.prompt.md` workflows\n- Workspace context and implementation constraints\n\n## Outputs\n\n- A conversion and enhancement plan\n- A safety and quality review checklist\n- File references and integration points\n- An implementation guide with verification steps\n\n## Rules\n\n> Core rules: [`templates/rules-core.md`](templates/rules-core.md)\n\n1. Run `context-map` before any conversion or planning work.\n2. Keep the pipeline idempotent and safe to re-run.\n3. Preserve the original task intent while improving clarity and structure.\n4. Keep manual safety notes unless they are wrong.\n5. Use git for rollback; do not create backup copies.\n6. Update the implementation plan when scope or requirements change.\n7. Prefer file-backed evidence over inference.\n\n## Skills Required\n\n> See full table with per-domain purposes:\n> [`templates/skills-table-core.md`](templates/skills-table-core.md)| Skill | Purpose |\n| ------------------------------------- | ---------------------------------------------------------------- |\n| `context-map` (prompt) | Map source, destination, and dependency impact before conversion |\n| `convert-plaintext-to-md` | Convert plaintext prompts to markdown |\n| `boost-prompt` | Improve prompt quality and structure |\n| `ai-prompt-engineering-safety-review` | Review prompts for safety and clarity |\n| `update-implementation-plan` | Keep the implementation plan current |\n| `prompt-builder` | Scaffold new prompt files |\n| `brainstorming` | Explore conversion and enhancement approaches |\n| `plans-and-specs` | Create detailed specs with file references |\n| `writing-skills` | Craft and improve prompts and instructions |\n| `writing-plans` | Structure the plan and its sections |\n| `acpx-executor` | Execute a prompt via any ACPX provider |\n| `executing-plans` | Execute implementation steps in order |\n| `simplify` | Keep the output concise and non-redundant |\n\n\n### Phase 1: Discover and verify\n\nConfirm the source files, target prompts, and required templates before planningchanges.\n\n### Phase 2: Plan conversion\n\nDefine how `.github/prompts/*.txt` should convert to `.github/prompts/*.md` and what must bepreserved.\n\n### Phase 3: Plan audits and fixes\n\nSpecify how to audit `boost-prompt`, `ai-prompt-engineering-safety-review`,`update-implementation-plan`, and `prompt-builder` with concrete findings.\n\n### Phase 4: Review and hand off\n\nProduce a compact completion report with execution steps, risk notes, and exitcriteria.\n\n### Phase 5: Execute Implementation Plan\n\nRun the conversion and enhancement pipeline end-to-end. Idempotent — safe to re-run.\n\n## Steps\n\n1. Load `context-map`, `brainstorming`, and `plans-and-specs`.\n2. Verify the conversion and audit templates exist.\n3. Map the prompt-library dependencies.\n4. Plan `.github/prompts/*.txt` → `.github/prompts/*.md` conversion.\n5. Plan audits for the related `.github/prompts/*.prompt.md` files.\n6. Confirm the target prompts reference `context-map` where relevant.\n7. Produce the final plan and handoff notes.\n\n## Tasks\n\n- [ ] Read the source inventory and target prompt files- [ ] Confirm the conversion and audit templates exist- [ ] Define the TXT→MD conversion scope and idempotency rules- [ ] Capture the feature specs, file references, and safety checks- [ ] List the prompts that need audit and cross-reference updates- [ ] Document the risk items and missing dependency links- [ ] Compile a concise completion report for implementation\n\n## Subtasks\n\n- [ ] Map `.github/prompts/*.txt` to their `.md` counterparts- [ ] Preserve source intent and safety notes in each converted file- [ ] Check for `context-map` references in each audit target- [ ] Record any missing template or format assumptions\n\n## Actions\n\n- `read_file(\"prompts/context-map.prompt.md\")` — Load the dependency map before planning\n- `skill_view(name=\"brainstorming\")` — Expand workflow options and tradeoffs\n- `skill_view(name=\"plans-and-specs\")` — Produce the detailed plan artifact\n- `skill_view(name=\"acpx-executor\")` — Dispatch tasks to ACPX providers\n- `search_files(pattern, target)` — Locate related prompt files and templates\n- `delegate_task(goal, toolsets)` — Split audit work when multiple files are involved---\n\n## Reference: `/enhance-markdown` Prompt\n\n> Full definition: `_archive/dev-init.prompts.txt` (lines 99–836)The `/enhance-markdown` prompt is the four-phase markdown auditor and enhancerused by the TXT→MD conversion pipeline. Key properties:```bash/enhance-markdown <file\n> [slug] # audit + enhance mode/enhance-markdown --txt-to-md [file] # TXT→MD conversion mode (batch or single)```- **9-section template**: Skills → Subagents → Personas → Rules → Phases → Steps → Tasks → Subtasks → Actions- **Batch size**: exactly 7 files per batch- **Resumable**: each phase checks for existing artifacts before re-running- **Plugin planning**: uses `createPlan`/`createSpec`/`appendSpec` with companion markdown fallback---\n\n## Phase 5: Execute Implementation Plan\n\n> Run the conversion and enhancement pipeline end-to-end. Idempotent — safe to re-run.\n\n### Entry Check\n\n```\n\nIF docs/prompt-conversion-enhancement-plan.md EXISTS → read it, skip to Step 5.3ELSE IF docs/dev-init-comprehensive-plan.md EXISTS → plan exists, skip to Step 5.2ELSE → run Phase 1–4 first, then return here\n```\n\n### Step 5.1 — Load Plan Artifacts\n\n```\n\nbashread_file(\"docs/dev-init-comprehensive-plan.md\")read_file(\"docs/dev-init-spec.md\")read_file(\"projects/Bash/archive/artifacts/context-maps/dev-init.context.json\")\n```\n\n### Step 5.2 — Run `context-map`\n\n1. Load `prompts/context-map.prompt.md`2. Map all `.github/prompts/*.txt` → `.github/prompts/*.md` pairs3. Map all `.github/prompts/*.prompt.md` dependencies4. Write `docs/prompts-cross-reference-registry.md`\n\n### Step 5.3 — Run `convert-plaintext-to-md`\n\nFor each `.github/prompts/*.txt` file:1. Read raw `.txt` content2. Apply enhancement (Stanford/Anthropic patterns): - Critical rules in first 15% - Nesting depth ≤ 4 - Instruction ratio 40–50% - Single source of truth (no rule repeated\n\n> 2×) - 3-tier prioritization: Safety → Core Workflow → Optimization3. Apply 9-section template: `Skills → Subagents → Personas → Rules → Phases → Steps → Tasks → Subtasks → Actions`4. Write to `.github/prompts/<same-stem>.md` (overwrite if exists)\n\n### Step 5.4 — Run `boost-prompt`\n\nFor each `.github/prompts/*.md` file:1. Load `prompts/boost-prompt.prompt.md`2. Apply quality enhancements: - Strengthen rule language (imperative, specific) - Add missing frontmatter (title, description, tags, trigger) - Ensure consistent heading hierarchy - Remove redundant sections3. **Constraint Preservation Audit**: log any rule removals with justification4. Write enhanced file back\n\n### Step 5.5 — Run `ai-prompt-engineering-safety-review`\n\nFor each `.github/prompts/*.md` and `.github/prompts/*.prompt.md`:1. Load `prompts/ai-prompt-engineering-safety-review.prompt.md`2. Check for: - Credential handling safety - Backup/rollback instructions present - Approval workflows for destructive ops - No fabricated verification claims3. Add safety notes where missing4. **Fail if critical safety constraints removed** — halt and report\n\n### Step 5.6 — Run `prompt-builder`\n\nFor any missing prompts:1. Load `prompts/prompt-builder.prompt.md`2. Scaffold new `.prompt.md` files from template3. Populate with content from plan artifacts\n\n### Step 5.7 — Run `update-implementation-plan`\n\n1. Load `prompts/update-implementation-plan.prompt.md`2. Update `docs/dev-init-comprehensive-plan.md` with: - Actual files processed - Issues found and resolved - Remaining items3. Mark completed phases\n\n### Tasks\n\n- [ ] Load all plan artifacts (Step 5.1)- [ ] Run context-map and write cross-reference registry (Step 5.2)- [ ] Convert all `.github/prompts/*.txt` → `.github/prompts/*.md` (Step 5.3)- [ ] Boost all `.github/prompts/*.md` files (Step 5.4)- [ ] Safety-review all prompts (Step 5.5)- [ ] Build missing prompts (Step 5.6)- [ ] Update implementation plan (Step 5.7)---\n\n## Phase 6: Verify All Prompts Optimal\n\n> Independent verification — read Phase 1 outputs only, not Phase 5 outputs (prevents confirmation bias).\n\n### Step 6.1 — Structural Audit\n\nFor each `.github/prompts/*.md`:\n\n```bash\n\n## Check frontmatterhead -10 Prompts/<name\n\n> .prompts.md | grep \"^title:\"head -10 Prompts/<name>.prompts.md | grep \"^description:\"head -10 Prompts/<name>.prompts.md | grep \"^trigger:\"head -10 Prompts/<name>.prompts.md | grep \"^tags:\"# Check required sections existgrep -c \"^\n\n## Steps\n\n\" Prompts/<name>.prompts.mdgrep -c \"^\n\n## Tasks\" Prompts/<name\n\n> .prompts.mdgrep -c \"^\n\n## Actions\n\n\" Prompts/<name>.prompts.md\n```\n\n### Step 6.2 — Cross-Reference Validation\n\n```bash\n\n## Verify all internal references resolvegrep -o '\\[.*\\](\\./[^)]*)' Prompts/<name>.prompts.md | while read ref; do path=$(echo \"$ref\" | sed 's/.*(\$.*\$)/\\1/') test -f \"$path\" || echo \"BROKEN REF: $ref in Prompts/<name>.prompts.md\"done\n\n```\n\n## Step 6.3 — Conflict Detection\n\n```bash\n\n## Check for duplicate triggers across Prompts/grep -h \"^trigger:\" Prompts/*.prompts.md | sort | uniq -d# Check for duplicate titlesgrep -h \"^title:\" Prompts/*.prompts.md | sort | uniq -d\n\n```\n\n## Step 6.4 — Quality Scoring\n\nScore each `.github/prompts/*.md`:| Criterion | Points | Detection || ------------------------------------\n\n- | ------ | ------------------------------------------------ || YAML frontmatter complete | +20 | title + description + trigger + tags present || Summary paragraph present | +15 | First paragraph after frontmatter is a summary || All required sections | +20 | Goal, Phases, Steps, Tasks, Actions present || Skills section references real skills | +10 | Each skill listed exists in `hermes skills list` || Actions use real tools | +10 | Each action maps to an available tool || No broken internal refs | +15 | All relative paths resolve || Consistent heading hierarchy | +10 | H1→H2→H3, no skipped levels |**Threshold**: ≥80 = optimal, 60–79 = needs work, <60 = rewrite required.\n\n### Step 6.5 — Write Verification Report\n\n```markdown\n\n## docs/dev-init-verification-report.md\n\n## Summary\n\n- Total .md files: N- Optimal (≥80): N- Needs work (60–79): N- Rewrite required (<60): N\n\n## Per-File Scores\n\n| File | Score | Issues || ---- | ----- | ------ || ... | ... | ... |\n\n## Conflicts Found\n\n| Type | Files | Resolution || ----------------\n\n- | ----- | ---------- || Duplicate trigger | ... | ... |\n\n## Broken References\n\n| File | Reference | Status || ---\n\n- | --------- | ------ || ... | ... | ... |\n```\n\n### Tasks\n\n- [ ] Structural audit all `.github/prompts/*.md` files- [ ] Validate all cross-references resolve- [ ] Detect duplicate triggers/titles- [ ] Score each file against quality criteria- [ ] Write verification report to `docs/dev-init-verification-report.md`- [ ] Fix any files scoring <80- [ ] Re-verify after fixes---\n\n## Phase 7: Cross-Reference Registry\n\n> Build the master registry of all prompts and their relationships.\n\n### Step 7.1 — Build Registry\n\nWrite `docs/prompts-cross-reference-registry.md`:```markdown# Prompts Cross-Reference Registry\n\n## Prompts/\\*.md (Conversion Targets)\n\n| File | Trigger | Status | Score | Depends On || --------------------\n\n- | ----------- | ------ | ----- | ------------------------------------ || skills-fix.prompts.md | /skills-fix | ✅ | 85 | context-map, skill-judge || dev-init.prompts.md | /dev-init | ✅ | 90 | context-map, convert-plaintext-to-md || ... | ... | ... | ... | ... |\n\n## prompts/\\*.prompt.md (Workflow Prompts)\n\n| File | Purpose | Used By || --------------------------------\n\n- | ------------------- | ------------------- || context-map.prompt.md | Dependency mapping | All dev-init phases || convert-plaintext-to-md.prompt.md | TXT→MD conversion | Phase 5.3 || boost-prompt.prompt.md | Quality enhancement | Phase 5.4 || ... | ... | ... |\n\n## Conflict Log\n\n| Conflict | Files | Resolution || -------\n\n- | ----- | ---------- || ... | ... | ... |```\n\n### Step 7.2 — Final Verification\n\n```\n\nbash# Count total promptsls Prompts/*.prompts.md | wc -lls prompts/*.prompt.md | wc -l# Verify registry matches disk# Every file on disk is in registry# Every registry entry exists on disk\n\n```\n\n### Tasks\n\n- [ ] Build cross-reference registry- [ ] Verify registry matches disk state- [ ] Commit all changes with `git commit -m \"dev-init: complete prompt-library refresh\"`---\n\n## Execution Summary\n\nWhen all 7 phases are complete, output:```========================================DEV INIT — PROMPT LIBRARY REFRESH COMPLETE========================================Phases executed: 7/7Prompts converted: N/N (.txt → .md)Prompts boosted: N/NSafety reviews: N passed, N failedQuality score ≥80: N/NConflicts resolved: NBroken refs fixed: NRegistry: docs/prompts-cross-reference-registry.mdVerification: docs/dev-init-verification-report.md===", "variables": [], "notes": "Create a complete plan for converting plaintext prompts to markdown, enhancing prompt quality, and updating related implementation plans.", "metadata": { "source": { "provider": "github-prompts", "repository": "https://github.com/rhixecompany/sandbox", "path": ".github/prompts/development/dev-init/dev-init.prompt.md", "ref": "a392d514ee6d8ff4ec01e32492d3aa7cc4601199", "url": "https://github.com/rhixecompany/sandbox/blob/a392d514ee6d8ff4ec01e32492d3aa7cc4601199/.github/prompts/development/dev-init/dev-init.prompt.md", "key": "rhixecompany/sandbox/.github/prompts/development/dev-init/dev-init.prompt.md" } } }

About