agentleFS
paruffparuff/uFawkesAI/.agents/skills/ai-stance/SKILL.md

ai-stance

Skill2 starsChanged 25 days agoMITView on GitHub

Write and maintain AI_STANCE.md: what AI use is expected, which tools are permitted, and what is prohibited, permitted with guardrails, or allowed. Use when onboarding a repo, adopting a new AI tool or model, or at the quarterly review. Implements DORA AI Capability 1.

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 1 is distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    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
  1. Very specific:

    Ask for: the repo name, any repo-specific prohibited uses, any tools beyond the defaults, and compliance needs (SOC2, GDPR).

    line 27
  2. Very specific:

    Run bash scripts/check-ai-stance.sh, then check by hand:

    line 47
  3. Distinctive:

    Every listed tool is in use, none is end-of-life, and each model id is current.

    line 49
  4. Very specific:

    When one uFawkes repo changes a shared item (a tool, a model, a Prohibited entry), compare the others: diff <(sed -n '/^## Permitted …

    line 57

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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

Skill: ai-stance

Load trigger: "load ai-stance skill" · DORA: AI Capability 1, a clear and communicated AI stance

AI_STANCE.md states how AI is used in the repo. scripts/check-ai-stance.sh (pre-commit and CI) enforces the structure: three policy buckets and a Last reviewed: date. This skill covers what the script cannot judge.

When to use

  • Onboarding a repo, or adopting a new AI tool or model (update the stance in the same PR).
  • The quarterly review, or after a reported violation of a Prohibited item.

Write it

Ask for: the repo name, any repo-specific prohibited uses, any tools beyond the defaults, and compliance needs (SOC2, GDPR). Then write AI_STANCE.md with these parts:

  1. Header. Last reviewed: <today>, Next review due: <today + 90 days>, owner.
  2. Expectation of use. Whether AI use is expected, encouraged, optional or discouraged, in one sentence of why. State that it applies to human contributors and AI agents.
  3. Organizational support. Permitted tools (each with a named model or version, not "latest LLM"), where the skills live, and how to raise a policy concern (an issue labelled ai-policy).
  4. Three buckets.
    • Prohibited (at least three items): sending PII, credentials or private infrastructure config to public models; committing AI-generated code that fails pre-commit; bypassing branch protection; unreviewed AI-generated security policy. Add repo-specific items, such as alerting rules or access policies that page people.
    • Permitted with guardrails: a table of use and guardrail, for example "AI code merged to main: human review and a test covering the change".
    • Allowed: genuinely low-risk uses such as completion, explanations, first drafts, triage.
  5. Agent obligations. Agents verify this file exists before work, log the session, and halt on a Prohibited action until a human authorizes it.

Audit it

Run bash scripts/check-ai-stance.sh, then check by hand:

  • Every listed tool is in use, none is end-of-life, and each model id is current.
  • Nothing in Allowed should have a guardrail; no "it depends" without saying on what.
  • Last reviewed is present and Next review due is not past; no [PLACEHOLDER] text remains.

Report gaps as Critical (block release), Non-critical (fix within 30 days), or Informational, each with the fix.

Keep it consistent across repos

When one uFawkes repo changes a shared item (a tool, a model, a Prohibited entry), compare the others: diff <(sed -n '/^## Permitted Tools/,/^## /p' ../<repo>/AI_STANCE.md) <(sed -n '/^## Permitted Tools/,/^## /p' AI_STANCE.md). Differences are fine when justified by the repo; unexplained ones are drift.

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.