agentleFS
Sign inStart free
z3e8z3e8/pyharness/.claude/skills/docs/SKILL.md

docs

Skill0 starsChanged 4 months agoApache-2.0View on GitHub

How to read, write, and maintain the pyharness docs under docs/. Use when adding or editing a docs page, deciding where documentation belongs, syncing docs after a code change, or when you need to consult the docs to understand a subsystem. Covers the three-section layout, the reference↔source map, and the "cut rather than rot" rule.

  • Reads credentials

165 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use import-memory, perf and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • amirbironSkill

    import-memory

    Import a memory export from another AI assistant into Claude's memory — conversationally, additively, and with the content treated as data.

    Also in 3 projects that use this

  • MicrosoftSkill

    perf

    Speed and memory performance rules for Rust crates, the Node addon, webui-framework, and webui-delivery.

    Also in 4 projects that use this

  • diego-tobalinaSkill

    build-fix

    Fix build and compilation errors systematically. Run when build fails.

    Also in 3 projects that use this

  • JkudjoSkill

    incident

    Production incident response — structured triage, hypothesis tree, safe remediation, incident record

    Also in 4 projects that use this

  • vbrevikSkill

    work

    Use when the user says 'plan', 'todo', 'copy plan', 'append plan', 'resume plan', 'priorities', or 'what's next'.

    Also in 7 projects that use this

  • mbwsimsSkill

    impact

    This skill should be used when the user asks to "check the impact", "what will break if I change this", "blast radius", "impact analysis", "what depends on this", "who imports this", "is it safe to change this", mentions "/impact", or wants to understand the consequences of modifying a specific file or function before making changes.

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

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

  • 0

    Common

    Most of the 129 edited versions we read keep these.

  • 5

    Distinctive

    This author's own choices. Worth a look.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Not a public product, not newcomers who aren't here.

    line 9
  2. Very specific:

    Do not bulk-read docs/ to "get context." Load the single page tied to the subsystem you're touching:

    line 15
  3. Distinctive:

    Don't recreate a tutorial quadrant.

    line 46
  4. Distinctive:

    Concrete over aspirational: Document what exists, not what you might build.

    line 71
  5. Distinctive:

    Cut before you pad: If a page has thinned to a stub that just points elsewhere, delete it and repoint its inbound links.

    line 76
  6. Very specific:

    Add it to the section list in docs/index.md.

    line 83
  7. Distinctive:

    If it's reference, add its source to the sync map above.

    line 84

“Common” is measured: kept by at least 6 in 10 of the 129 edited versions of this file we could read. “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

Maintaining the pyharness docs

The docs exist for one audience: you and the agents working in this repo. Not a public product, not newcomers who aren't here. Every page must earn its keep as non-inferable knowledge someone will actually reach for. When in doubt, cut it — a missing page costs a grep; a stale page lies.

Read narrowly, not broadly

Do not bulk-read docs/ to "get context." Load the single page tied to the subsystem you're touching:

Working on…Read
the run_python model, kernel, delegationexplanation/action-space.md
dispatch / capability routingexplanation/broker.md
policy, vault, sandbox, audit chainexplanation/security-and-audit.md
budget, tiers, pricingexplanation/budget.md
the builtin surfacereference/builtins.md
env vars / configreference/configuration.md
the public Python APIreference/python-api.md
the CLIsreference/cli.md

The code is ground truth; a doc page is a faster, curated summary. If a page disagrees with the code, the code wins — and you fix the page (see below).

The three sections (Diátaxis, minus tutorials)

Each page stays in exactly one lane. Mixing lanes is the main way these rot.

  • explanation/ — why and how it works. Architecture and rationale. Most durable; changes only when the design changes. New conceptual background goes here.
  • how-to/ — one real task, start to finish. "Add a tool," "use the vault," "run with observability." If it's not a task someone actually performs, it doesn't belong. (This is why there's no deploy page — nothing is deployed.)
  • reference/ — precise lookup. Tables of signatures, flags, env vars. Lean; no narrative. This section drifts fastest — see the sync map.

There is no tutorials/ section. Onboarding lives in the top-level README.md. Don't recreate a tutorial quadrant.

Sync docs after a code change (do this in the same commit)

A change is not done until the docs it affects are true again. The reference pages track specific sources — when you touch the left, update the right:

If you change…Update
the SYSTEM_PROMPT builtins in pyharness/core/agent.pyreference/builtins.md
env vars (add/rename/default) in .env.examplereference/configuration.md
the public exports in pyharness/__init__.pyreference/python-api.md
CLI flags/behavior in cli/main.py / cli/vault.pyreference/cli.md
tiers, pricing, or budget enforcementexplanation/budget.md
policy defaults, vault rules, or the sandboxexplanation/security-and-audit.md
how side effects route through the brokerexplanation/broker.md

Also check README.md (it has the quickstart and a few illustrative builtin names, then links to reference/builtins.md for the full set) and the repo map in AGENTS.md when structure moves.

Writing rules

  • Lean. Reference is tables, not prose. Link instead of repeating — each fact has one home; other pages point to it.
  • Concrete over aspirational. Document what exists, not what you might build. No "in a shared deployment you would…" pages.
  • Real internal links. Paths are relative to the page's own directory; the cross-links between pages are load-bearing, so don't break them when moving content.
  • Cut before you pad. If a page has thinned to a stub that just points elsewhere, delete it and repoint its inbound links.

When adding a genuinely new page

  1. Confirm it's non-inferable and someone will reach for it. If not, don't.
  2. Pick the one lane it belongs to; if it spans two, it's two pages.
  3. Add it to the section list in docs/index.md.
  4. If it's reference, add its source to the sync map above.

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