AGENTS.md

AGENTS

• Read the repo profile selector at .matrixai/repo-profile.yml.
• Enforce the universal hotset .matrixai/matrixai-standards/standards/HOTSET.md.
• Enforce the profile doc under .matrixai/matrixai-standards/standards/profiles/ matching profile: in .matrixai/repo-profile.yml (e.g., library-js, application-js, worker-js-cloudflare, docusaurus-js-cloudflare).
• Profile index (for discovery): .matrixai/matrixai-standards/standards/profiles/README.md
• Tooling contract reference: .matrixai/matrixai-standards/standards/coding/tooling/tooling-contract.md (commands, artifacts, expectations per profile).
• Prefer ASCII punctuation/symbols when an equivalent exists (see .matrixai/matrixai-standards/standards/HOTSET.md [MXS-GEN-006]).
• Ensure edits comply with .editorconfig (LF line endings, 2-space indent, final newline, trim trailing whitespace).
• Line-reference policy (applies to all agent-generated repository content: Markdown, docs, templates, and code comments):
  • Never emit path:line (e.g. foo.ts:1, README.md:126) into repository files.
  • Do NOT put :number inside Markdown link destinations: [x](path:123) is banned.
  • If a line reference is needed, use either:
    • [x](path#heading-anchor) (if possible), or
    • [x](path) (line 123) (preferred, portable), or
    • [x](path#L123) only when explicitly targeting a renderer that supports #L anchors.
  • If you would have emitted :1, drop it entirely: use path with no line info.
• Apply repo-local golden commands and overrides here (use npm run lintfix during active development; use npm run lint for non-mutating CI checks):
  • build: npm run build
  • test: npm test
  • lintfix: npm run lintfix
  • lint: npm run lint
  • docs: npm run docs
  • bench: (not defined in this repo)