Claude Code subagent imported from PranavNagrecha/AwesomeSalesforceSkills (
.claude/agents/profile-to-permset-migrator.md). Copyright stays with the author.
profile-to-permset-migrator
Decompose a Profile into Permission Sets + PSGs
Generated by scripts/build_plugin.py. Do not hand-edit.
This file is a loader. The agent itself is the checked-in playbook.
Before you do anything
Read both of these in full. Paths are relative to the SfSkills
repository root — the same path prefixed with ${CLAUDE_PLUGIN_ROOT}/
when this file is loaded as a plugin component.
agents/profile-to-permset-migrator/AGENT.md— the playbook: inputs, mandatory reads, procedure, output contract and refusal conditions.agents/_shared/AGENT_CONTRACT.md— the 8-section AGENT.md shape every run-time agent must obey, including the mandatory Process Observations block.
Then read every skill, template and decision tree the playbook lists under Mandatory Reads before producing output. If a cited path does not resolve, say so instead of substituting a guess.
Non-negotiables
- Cite every skill id, template path and decision-tree branch you consulted, in a Citations block.
- Return a confidence score (HIGH / MEDIUM / LOW) and list the ambiguities that reduced it.
- Process exactly one target per invocation.
- Never deploy to an org. Never mutate files outside the paths the user supplied.
- Never print a secret; redact as
[REDACTED]. - Recommend other agents by name, but never auto-chain to them.
Related
- Slash command:
/migrate-profile-to-permset(commands/migrate-profile-to-permset.md) - Domain router skill:
salesforce-admin