agentleFS
dachentdachent/skills/repo-map-codex/SKILL.md

repo-map-codex

Skill27 starsChanged 3 months agoMITView on GitHub

Deprecated for GPT-5.6 Sol. Native Codex Plan Mode performs read-first repository grounding and evidence collection.

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 2 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Create a durable map of the project so later planning is grounded in evidence rather than memory.

    line 12
  2. Very specific:

    Use the orchestrator-selected project mode unless discovered evidence contradicts it.

    line 24
  3. Very specific:

    Write .deep-planning/repo-map.md and .deep-planning/evidence-catalog.md using the deep-planning artifact contract.

    line 25
  4. Very specific:

    Use N/A in Failure mechanism for successful evidence rows.

    line 26
  5. Very specific:

    If required evidence is missing, set status to BLOCKED_BY_MISSING_EVIDENCE and make Next action name the exact source, access, or probe …

    line 28
  6. Distinctive:

    Read the relevant reference before writing the map:

    line 32
  7. Very specific:

    When used inside deep planning, write only to the target project's .deep-planning/repo-map.md and .deep-planning/evidence-catalog.md unless …

    line 56

≈ “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

Codex Repo Map

Deprecated for GPT-5.6 Sol. Native Codex Plan Mode performs read-first repository grounding and evidence collection. Do not use this package for new work.

Overview

Create a durable map of the project so later planning is grounded in evidence rather than memory. The map adapts to software with Git, software without Git, business artifact work, and mixed business-coding projects.

Workflow

  1. Determine the project mode: software-git, software-no-git, business-artifact, or mixed-business-coding.
  2. Explore read-only first. Prefer rg --files, rg, normal file reads, and parallel reads with multi_tool_use.parallel when available.
  3. Build a concise repo-map.md or equivalent project map.
  4. Build an evidence-catalog.md that lists source materials, attempt/outcome status, and relevance.
  5. Record unknowns separately from facts. Do not ask the user for facts that can be discovered.

When Invoked By deep-planning-codex

  • Use the orchestrator-selected project mode unless discovered evidence contradicts it. If the mode changes, record the reason in .deep-planning/project-mode.md.
  • Write .deep-planning/repo-map.md and .deep-planning/evidence-catalog.md using the deep-planning artifact contract. The evidence catalog must include Failure mechanism for failed, partial, stale, or unreached evidence.
  • Use N/A in Failure mechanism for successful evidence rows.
  • End with a Deep Planning Delta for the orchestrator to merge into .deep-planning/state.md. Use READY_FOR_PROCEED only when the map and evidence catalog are complete enough for the next planning phase.
  • If required evidence is missing, set status to BLOCKED_BY_MISSING_EVIDENCE and make Next action name the exact source, access, or probe needed.

Mode References

Read the relevant reference before writing the map:

  • references/software-git.md
  • references/software-no-git.md
  • references/business-artifact.md
  • references/mixed-business-coding.md

Minimum Output

The map must include:

  • project mode and why it was selected;
  • key source materials;
  • entrypoints or primary deliverables;
  • relevant commands, if any;
  • source-of-truth data/docs;
  • dependencies and external systems;
  • generated or derived artifacts;
  • validation surfaces;
  • risk areas;
  • unknowns that require user or stakeholder input.

Write Scope

When used inside deep planning, write only to the target project's .deep-planning/repo-map.md and .deep-planning/evidence-catalog.md unless the user specifies another planning workfolder.

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.