agentleFS
IgorGanapolskyIgorGanapolsky/mac-yolo-safeguards/.github/copilot-instructions.md

mac-yolo-safeguards

Copilot instructions3 starsChanged 3 days agoMITView on GitHub

Canonical directive for all GitHub Copilot Agents, Copilot Workspace sessions, and automated PR review agents operating on this repository. Every code modification must pass verification before opening or merging a PR: - Unit & System Verification: bash scripts/verify.sh - Mobile Unit Tests: npm test -- --no-coverage --watchman=false (under /hermes-mobile) - CodeQL Security: Zero new CodeQL alerts on main. When creating or reviewing Pull Requests: 1. Title Format: Conventional commits (feat(...), fix(...), ci(...), docs(...)). 2. Body Metadata: Include Linear ticket reference…

The file

We split this copilot-instructions.md into the rules it gives an agent

8 rules. About 0 are common, 6 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

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

    Multi-Agent Coordination: Always respect repo locks in plan.md and Linear issue locks (tools/linear-agent-bridge.js).

    line 9
  2. Distinctive:

    No Desktop Hijack: Never launch headed Chrome, hijack macOS GUI focus, or touch Igor's daily profile unless explicitly requested in that …

    line 11
  3. Distinctive:

    Every code modification must pass verification before opening or merging a PR:

    line 18
  4. Distinctive:

    When creating or reviewing Pull Requests:

    line 27
  5. Distinctive:

    Auto-Merge Tagging: If PR passes all CI checks and contains non-breaking code, enable auto-merge:

    line 30
  6. Distinctive:

    Fresh-User Onboarding Mindset: Treat every test as a brand-new user (no saved profiles, release build).

    line 53
  7. Distinctive:

    Never inline, print, or commit API keys, PATs, or secrets.

    line 61
  8. Very specific:

    Use macOS Keychain / local authenticated stores (gh auth status).

    line 62

≈ “Common” is an estimate: at least 1 in 5 of the 3,827 Copilot instructions 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
copilot-instructions.md

GitHub Copilot Agent & Workspace Instructions

Canonical directive for all GitHub Copilot Agents, Copilot Workspace sessions, and automated PR review agents operating on this repository.


1. Core Operating Directives

  • Multi-Agent Coordination: Always respect repo locks in plan.md and Linear issue locks (tools/linear-agent-bridge.js). Never edit files claimed by another active agent.
  • Always Agent Mode: Execute read-write and command tasks autonomously; never ask the user to run manual commands.
  • No Desktop Hijack: Never launch headed Chrome, hijack macOS GUI focus, or touch Igor's daily profile unless explicitly requested in that message.
  • Zero Approval Debt: Small, reviewable PR slices. Always run verification commands before declaring completion.

2. Verification & Quality Gates

Every code modification must pass verification before opening or merging a PR:

  • Unit & System Verification: bash scripts/verify.sh
  • Mobile Unit Tests: npm test -- --no-coverage --watchman=false (under /hermes-mobile)
  • CodeQL Security: Zero new CodeQL alerts on main.

3. PR Creation & Speed Optimization

When creating or reviewing Pull Requests:

  1. Title Format: Conventional commits (feat(...), fix(...), ci(...), docs(...)).
  2. Body Metadata: Include Linear ticket reference (Closes AGENT-XX), files modified, and explicit verification output (commit SHA, test run results).
  3. Auto-Merge Tagging: If PR passes all CI checks and contains non-breaking code, enable auto-merge:
    gh pr merge <PR_NUMBER> --squash --auto
  4. Copilot code review on every PR: Request an automated Copilot review before human review to catch regressions early and shrink review cycles.
    # via gh (requires you in CODEOWNERS/reviewers): tag Copilot as a reviewer
    gh pr edit <PR_NUMBER> --add-co "Copilot"
    # or leave a PR comment: @copilot review
  5. Copilot Autfix for security alerts (Aug 2026): For any new CodeQL/secret-scanning/dependency alert this PR introduces, open an autofix branch and reference it in the body. Track under § Risk.
  6. Agentic reply iteration: Address review comments by replying @copilot fix / @copilot apply on the thread instead of hand-rolling — keeps cycles to a single touch.
  7. Merge queue + auto-rebase (kills #1407 conflict class): On protected branches, require the PR to be up to date with base before merge (or route through a merge queue) so lingering branches never re-conflict. Config one-liner (repo-admin, needs explicit go):
    gh api repos/$(gh repo view --json nameWithOwner -q .nameWithOwner)/branches/main/protection \
      --method PUT --input scripts/merge-queue-protection.json
  8. Attribution & Telemetry: Attach agent labels (agent:copilot, agent:<primary>) and cycle-time metadata.

4. Hermes Mobile Specific Rules

  • Fresh-User Onboarding Mindset: Treat every test as a brand-new user (no saved profiles, release build).
  • Expo SDK Pins Are Law: react-native, react, and expo* dependencies must move only via npx expo install --fix.
  • Maestro E2E Input Protocol: Chat input testing must use ONLY "make money today".

5. Security & Credentials

  • Never inline, print, or commit API keys, PATs, or secrets.
  • Use macOS Keychain / local authenticated stores (gh auth status).

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.