How long does the integration take?

Change one base URL in an OpenAI-compatible client, or install the Python or TypeScript SDK; the quickstart routes a first request in a few minutes.

Will this lock us into another vendor?

No. The router is Apache-2.0 open source, your provider keys stay yours, and the API is OpenAI-compatible, so leaving is a change of base URL.

What happens when a provider fails?

The router falls back to the next model your policy allows and records the fallback in the trace, so an outage costs a retry, not an incident.

trellis-finish-work - Prompt

Name: trellis-finish-work
Author: castbox

Prompt file imported from castbox/guru-trellis (.cursor/commands/trellis-finish-work.md). Copyright stays with the author.

Finish Work

Active workflow ownership

Before formal work, load the validated Task workspace workflow with:

python3 .trellis/scripts/get_context.py --mode runtime-contract
python3 .trellis/scripts/get_context.py --mode continuation

A declared Delivery contract owns all formal routing. Stop on malformed or unsupported contracts. If the role owner differs from this Skill, hand off to that declared owner and stop; do not run a parallel native gate. Read .trellis/delivery-contract.md and the exact current Delivery query before recording results. missing_runtime_contract means the existing legacy workflow has no declaration; retain the legacy behavior below in that case. Task-free work creates no formal lifecycle result.

Delivery closeout profile

With a Delivery contract, defer to its declared closure owner. Consume an actual current-generation whole Task Completion result, then assess source disposition. Default is no_mutation. Issue closure requires separate authorization and live outcome. Record close_task through the declared owner, archive only the selected completed Task if authorized, and treat cleanup as a separate action. Archive never records completion; failed cleanup preserves it. Never demand a synthetic commit for manual/Git-free delivery. Return here only after trellis-complete-task has assessed all accepted scope, unallocated work and evidence. Stop after this profile; the legacy archive instructions below do not apply.

Wrap up the current session: archive the active task (and any other completed-but-unarchived tasks the user wants to clean up). Code commits are NOT done here — those happen in workflow Phase 3.4 before you invoke this command.

Step 1: Survey current state

python3 ./.trellis/scripts/get_context.py

Context distinguishes the current-session task from the project task inventory. Read the current task's acceptance criteria and verification evidence before archiving it. Other tasks require explicit selection and user confirmation; never infer personal task ownership.

Step 2: Sanity check — classify dirty paths

Run:

git status --porcelain

Classify task metadata changes separately from code; leave unrelated changes untouched.

For each remaining dirty path, decide whether it belongs to the current task or to other parallel work (e.g., another terminal window editing the same repo). Heuristics:

Paths referenced in the current task's prd.md / implement.jsonl / check.jsonl → current task
Paths in code areas matching the task's stated scope, or that you remember editing this session → current task
Paths in unrelated areas you have no recollection of touching this session → other parallel work

Then route:

Any remaining path looks like current-task work — bail out with:

"Working tree has uncommitted code changes from this task: <list>. Return to workflow Phase 3.4 to commit them before running /trellis-finish-work."

Do NOT run git commit here. Do NOT prompt the user to commit. The user goes back to Phase 3.4 and the AI drives the batched commit there.
All remaining paths look unrelated (other parallel-window work) — report them once and continue to Step 3:

"FYI, dirty files outside this task's scope — leaving them for the other window: <list>."
Genuinely unsure — ask the user once: "Are <list> this task's work I forgot to commit, or another window's? (commit / ignore)" — then route per their answer.

Step 3: Archive task(s)

python3 ./.trellis/scripts/task.py archive <task-name>

Archive the current task only when its acceptance criteria and verification evidence establish completion. If incomplete, leave it active and report remaining work. Additional tasks require the explicit confirmation from Step 1. Archive auto-commit defaults to true, controlled by task_auto_commit; honor explicit opt-outs and use --no-commit when commits are not authorized.

If there is no active task and the user did not confirm any cleanup archives, skip this step.

Step 4: Report outcome

Summarize the archived task, validation results, and remaining follow-ups. Keep durable evidence in the task artifacts; no additional recording step is needed.

Install trellis-finish-work and it becomes one more candidate the router can pick - when it fits.

A listing is a routing target with a manifest: what it does, which domains it covers, what it costs and who publishes it. Once installed it sits beside your own models and tools, is scored like any other candidate for each request, and shows up in the trace when it wins. Ratings come from workspaces that installed it, one per account.

# Finish Work ## Active workflow ownership Before formal work, load the validated Task workspace workflow with: ```bash python3 .trellis/scripts/get_context.py --mode runtime-contract python3 .trellis/scripts/get_context.py --mode continuation ``` A declared Delivery contract owns all formal routing. Stop on malformed or unsupported contracts. If the role owner differs from this Skill, hand off to that declared owner and stop; do not run a parallel native gate. Read `.trellis/delivery-contract.md` and the exact current Delivery query before recording results. `missing_runtime_contract` means the existing legacy workflow has no declaration; retain the legacy behavior below in that case. Task-free work creates no formal lifecycle result. ## Delivery closeout profile With a Delivery contract, defer to its declared closure owner. Consume an actual current-generation whole Task Completion result, then assess source disposition. Default is no_mutation. Issue closure requires separate authorization and live outcome. Record close_task through the declared owner, archive only the selected completed Task if authorized, and treat cleanup as a separate action. Archive never records completion; failed cleanup preserves it. Never demand a synthetic commit for manual/Git-free delivery. Return here only after trellis-complete-task has assessed all accepted scope, unallocated work and evidence. Stop after this profile; the legacy archive instructions below do not apply. Wrap up the current session: archive the active task (and any other completed-but-unarchived tasks the user wants to clean up). Code commits are NOT done here — those happen in workflow Phase 3.4 before you invoke this command. ## Step 1: Survey current state ```bash python3 ./.trellis/scripts/get_context.py ``` Context distinguishes the current-session task from the project task inventory. Read the current task's acceptance criteria and verification evidence before archiving it. Other tasks require explicit selection and user confirmation; never infer personal task ownership. ## Step 2: Sanity check — classify dirty paths Run: ```bash git status --porcelain ``` Classify task metadata changes separately from code; leave unrelated changes untouched. For each remaining dirty path, decide whether it belongs to **the current task** or to **other parallel work** (e.g., another terminal window editing the same repo). Heuristics: - Paths referenced in the current task's `prd.md` / `implement.jsonl` / `check.jsonl` → current task - Paths in code areas matching the task's stated scope, or that you remember editing this session → current task - Paths in unrelated areas you have no recollection of touching this session → other parallel work Then route: - **Any remaining path looks like current-task work** — bail out with: > "Working tree has uncommitted code changes from this task: `<list>`. Return to workflow Phase 3.4 to commit them before running `/trellis-finish-work`." Do NOT run `git commit` here. Do NOT prompt the user to commit. The user goes back to Phase 3.4 and the AI drives the batched commit there. - **All remaining paths look unrelated** (other parallel-window work) — report them once and continue to Step 3: > "FYI, dirty files outside this task's scope — leaving them for the other window: `<list>`." - **Genuinely unsure** — ask the user once: "Are `<list>` this task's work I forgot to commit, or another window's? (commit / ignore)" — then route per their answer. ## Step 3: Archive task(s) ```bash python3 ./.trellis/scripts/task.py archive <task-name> ``` Archive the current task only when its acceptance criteria and verification evidence establish completion. If incomplete, leave it active and report remaining work. Additional tasks require the explicit confirmation from Step 1. Archive auto-commit defaults to true, controlled by `task_auto_commit`; honor explicit opt-outs and use `--no-commit` when commits are not authorized. If there is no active task and the user did not confirm any cleanup archives, skip this step. ## Step 4: Report outcome Summarize the archived task, validation results, and remaining follow-ups. Keep durable evidence in the task artifacts; no additional recording step is needed.

{ "prompt": "# Finish Work\n\n## Active workflow ownership\n\nBefore formal work, load the validated Task workspace workflow with:\n\n```bash\npython3 .trellis/scripts/get_context.py --mode runtime-contract\npython3 .trellis/scripts/get_context.py --mode continuation\n```\n\nA declared Delivery contract owns all formal routing. Stop on malformed or\nunsupported contracts. If the role owner differs from this Skill, hand off to\nthat declared owner and stop; do not run a parallel native gate. Read\n`.trellis/delivery-contract.md` and the exact current Delivery query before\nrecording results. `missing_runtime_contract` means the existing legacy workflow has no declaration;\nretain the legacy behavior below in that case. Task-free work creates no formal lifecycle result.\n\n## Delivery closeout profile\n\nWith a Delivery contract, defer to its declared closure owner. Consume an actual\ncurrent-generation whole Task Completion result, then assess source disposition.\nDefault is no_mutation. Issue closure requires separate authorization and live\noutcome. Record close_task through the declared owner, archive only the selected\ncompleted Task if authorized, and treat cleanup as a separate action. Archive\nnever records completion; failed cleanup preserves it. Never demand a synthetic\ncommit for manual/Git-free delivery. Return here only after trellis-complete-task\nhas assessed all accepted scope, unallocated work and evidence. Stop after this\nprofile; the legacy archive instructions below do not apply.\n\nWrap up the current session: archive the active task (and any other completed-but-unarchived tasks the user wants to clean up). Code commits are NOT done here — those happen in workflow Phase 3.4 before you invoke this command.\n\n## Step 1: Survey current state\n\n```bash\npython3 ./.trellis/scripts/get_context.py\n```\n\nContext distinguishes the current-session task from the project task inventory. Read the current task's acceptance criteria and verification evidence before archiving it. Other tasks require explicit selection and user confirmation; never infer personal task ownership.\n\n## Step 2: Sanity check — classify dirty paths\n\nRun:\n\n```bash\ngit status --porcelain\n```\n\nClassify task metadata changes separately from code; leave unrelated changes untouched.\n\nFor each remaining dirty path, decide whether it belongs to **the current task** or to **other parallel work** (e.g., another terminal window editing the same repo). Heuristics:\n\n- Paths referenced in the current task's `prd.md` / `implement.jsonl` / `check.jsonl` → current task\n- Paths in code areas matching the task's stated scope, or that you remember editing this session → current task\n- Paths in unrelated areas you have no recollection of touching this session → other parallel work\n\nThen route:\n\n- **Any remaining path looks like current-task work** — bail out with:\n > \"Working tree has uncommitted code changes from this task: `<list>`. Return to workflow Phase 3.4 to commit them before running `/trellis-finish-work`.\"\n\n Do NOT run `git commit` here. Do NOT prompt the user to commit. The user goes back to Phase 3.4 and the AI drives the batched commit there.\n- **All remaining paths look unrelated** (other parallel-window work) — report them once and continue to Step 3:\n > \"FYI, dirty files outside this task's scope — leaving them for the other window: `<list>`.\"\n- **Genuinely unsure** — ask the user once: \"Are `<list>` this task's work I forgot to commit, or another window's? (commit / ignore)\" — then route per their answer.\n\n## Step 3: Archive task(s)\n\n```bash\npython3 ./.trellis/scripts/task.py archive <task-name>\n```\n\nArchive the current task only when its acceptance criteria and verification evidence establish completion. If incomplete, leave it active and report remaining work. Additional tasks require the explicit confirmation from Step 1. Archive auto-commit defaults to true, controlled by `task_auto_commit`; honor explicit opt-outs and use `--no-commit` when commits are not authorized.\n\nIf there is no active task and the user did not confirm any cleanup archives, skip this step.\n\n## Step 4: Report outcome\n\nSummarize the archived task, validation results, and remaining follow-ups. Keep durable evidence in the task artifacts; no additional recording step is needed.", "variables": [], "notes": "Before formal work, load the validated Task workspace workflow with:", "metadata": { "source": { "provider": "github-cursor-commands", "repository": "https://github.com/castbox/guru-trellis", "path": ".cursor/commands/trellis-finish-work.md", "ref": "ab854ef29448910c332faadeec3c1a5dd810901d", "url": "https://github.com/castbox/guru-trellis/blob/ab854ef29448910c332faadeec3c1a5dd810901d/.cursor/commands/trellis-finish-work.md", "key": "castbox/guru-trellis/.cursor/commands/trellis-finish-work.md" } } }

trellis-finish-work

About

Finish Work

Active workflow ownership

Delivery closeout profile

Step 1: Survey current state

Step 2: Sanity check — classify dirty paths

Step 3: Archive task(s)

Step 4: Report outcome

Install trellis-finish-work and it becomes one more candidate the router can pick - when it fits.

What it costs to adopt - in engineering time.

One base URL

No vendor lock-in

Failover built in

Use it

Manifest

Reviews