agentleFS
kaithoughtarchitectkaithoughtarchitect/toppingout/skills/f-context/SKILL.md

f-context

Skill14 starsChanged 20 days agoMITView on GitHub

Build a full orientation brief for a codebase you are arriving at cold, stack, git state, active work, focus-matched Logbook knowledge, relevant architecture, and standing warnings. EXPENSIVE: optional explorers; more with --deep. Use when the user runs /f-context, says "get me up to speed on this project", "what is this codebase", "orient me", or is resuming after days away. Not for a mid-session refresher, which is free (f-status). Not for Logbook card search (f-search).

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 2 are distinctive, and 1 is 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.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Do not re-explore facts already established by trustworthy evidence unless a contradiction appears.

    line 43
  2. Very specific:

    If a focus was given, add a short deep dive for that area.

    line 57
  3. Distinctive:

    Never edit files, stage, commit, push, call paid services, or perform live writes.

    line 61

≈ “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

Topping Out: Logbook context

This is a read-only orientation pass.

Usage

/f-context
/f-context [focus-area]
/f-context --deep [focus-area]

No arguments: Full project context.

--deep: Adds commit archaeology, symbol/impact tracing, pattern comparison, and file-neighborhood analysis. Use after a long break or before resuming complex in-flight work; skip for routine session starts.

With focus: Prioritize that area (a path, system name, or live Logbook cluster slug, do not invent a closed catalog).

Collection

Two waves. Identify active work first so Logbook search has a target.

Wave 1 (root): Read the constitution (the repo profile names it: AGENTS.md, or CLAUDE.md where that is the constitution), git status / recent commits / branch, the open job in foreman/jobs/ (if several, the newest by filename date unless the owner names one; skip any whose JOB COMPLETE (older records: COMMISSION COMPLETE) status reads ✅ COMPLETE, that is finished work awaiting /f-signoff), the repo profile's gates and the repo's own build/test entry points (whatever manifest the stack uses: package.json scripts, pyproject.toml, a Makefile, Cargo.toml, …), and top-level structure. Trace architecture, entry points, and tests for the current work area, not the whole repo. Write one focus line: "Active work: <topic/files/system>".

Wave 2 (root): Search current Logbook cards in foreman/logbook/entries/ against that focus line (focus-matched cards first). Count live cards ([A-Z]*_20*.md, excluding *TEMPLATE*). If the count is greater than 250, run /f-archive as a dry run only (this is a read-only orientation: it reports that an archive is due and changes nothing) before synthesizing, that preview is read-only. Read 2–3 recent completed jobs in foreman/completed-jobs/ that touch the focus when present.

Root may call code-explorer with bounded, non-overlapping questions. Load .claude/agents/code-explorer.md. Child roles return evidence only and never recursively dispatch.

In --deep mode, also: full recent-commit bodies for WHY; symbol/impact tracing on the most-changed files; similar implementations to follow; same-directory neighborhoods of dirty or recently-committed files.

Do not re-explore facts already established by trustworthy evidence unless a contradiction appears. The constitution and the open job may already be in context.

Output

Synthesize:

  • project snapshot and current priority;
  • dirty-tree and recent-commit state;
  • active work and ownership boundaries;
  • focus-matched Logbook knowledge;
  • relevant architecture and data flow;
  • critical warnings and environment constraints (from the constitution);
  • a short recommended starting sequence with source paths.

If a focus was given, add a short deep dive for that area. If --deep, add a Deep Context section (symbol map, blast radius, patterns-to-follow, neighborhoods, WHY from commit bodies). If a gather step fails, continue with what you have and say so.

Boundaries

Never edit files, stage, commit, push, call paid services, or perform live writes.

Cost is real. Mid-session "where are we" is /f-status and costs nothing. At a job's start, /f-shift already loads the job's cards; use this for arriving at a repo cold, or before re-scoping. Say so rather than running this.

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.