agentleFS
Sign inStart free
loganware05loganware05/captains-compass-cursor/.cursor/skills/project-overseer/SKILL.md

project-overseer

Skill2 starsChanged 33 days agoMITView on GitHub

Read-oriented Project Overseer — snapshot, explain, compare, and propose memory reconcile without mutating authority

  • Reads credentials

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 4 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  • could be automatic: a setting could enforce it
  1. Distinctive:

    proposed: Prefer citations to files, SHAs, VERSION, and evidence paths.

    line 58
  2. Very specific:

    Treat .agent/status/system-map.md and scope-ledger.json as Overseer artifacts — not replacements for PROJECT_CONTEXT.md / ADRs.

    line 60
  3. Very specific:

    Never approve IMPLEMENTATION_PLAN.md or set plan Status to APPROVED.

    line 65
  4. Distinctive:

    Never promote Skills or Policies.

    line 66
  5. Distinctive:

    Never merge, deploy, release, or rewrite canonical project memory silently.

    line 67
  6. Distinctive:

    Never require live HTTP/Kimi for CI.

    line 68
  7. Very specific:

    Always keep approved_for_execution: false and authority_mutation: false.

    line 70

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Project Overseer

Use this Skill when

The Captain or First Mate needs a current, evidence-backed understanding of repository reality vs project memory — drift detection, system-map refresh, or reconcile proposals — without granting execution authority.

Inputs

  • Repository root (--repo / --repo-root)
  • Optional enable gate: COMPASS_STATUS_ENABLED=1
  • Optional provider: COMPASS_STATUS_PROVIDER=stub|http (default stub)
  • Optional live model: COMPASS_STATUS_MODEL + OPENROUTER_API_KEY / COMPASS_STATUS_API_KEY when provider is http

Procedure

  1. Enable the gate (required):

    export COMPASS_STATUS_ENABLED=1

  2. Snapshot repository understanding:

    ./scripts/northstar status snapshot --repo .

  3. Explain (stub by default; opt-in live):

    ./scripts/northstar status explain --repo .
# Optional live path (never required for CI):
COMPASS_STATUS_PROVIDER=http COMPASS_STATUS_MODEL=<model> \
  ./scripts/northstar status explain --repo .

  4. Compare since a SHA:

    ./scripts/northstar status compare --since <sha> --repo .

  5. Reconcile (proposal-only drafts under .agent/status/reconcile-proposals/):

    ./scripts/northstar status reconcile --repo .
  6. Label every material claim as observed, documented, external, or proposed. Prefer citations to files, SHAs, VERSION, and evidence paths.
  7. Treat .agent/status/system-map.md and scope-ledger.json as Overseer artifacts — not replacements for PROJECT_CONTEXT.md / ADRs.

Constraints

  • Never approve IMPLEMENTATION_PLAN.md or set plan Status to APPROVED.
  • Never promote Skills or Policies; never mutate routing or reputation.
  • Never merge, deploy, release, or rewrite canonical project memory silently.
  • Never require live HTTP/Kimi for CI; stub must remain sufficient.
  • Fail closed when the live provider is selected without keys or on network error.
  • Always keep approved_for_execution: false and authority_mutation: false.

Return

  • Snapshot / explanation summary with claim labels
  • Paths to system-map, scope-ledger, and any reconcile proposals
  • Explicit statement that canonical memory was not auto-rewritten

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free