| name | harness-lint |
| description | Use when working in this repository after receiving recurring coding feedback, adding or reviewing durable lint guardrails from AGENTS.md, updating rules/*.md, or finishing changes that should be checked by harness-lint. |
Harness Lint
Use harness-lint to turn recurring repository feedback into executable or reviewable rules.
Workflow
- Run commands from the repository root.
- Use
harness-lint from PATH. If it is missing on this machine and /Users/ctrdh/Code/harness-lint/target/debug/harness-lint exists, that local debug binary is an acceptable fallback.
- When the user gives a recurring coding preference, create or update a rule under
rules/ instead of only changing the current code.
- Rules must use GritQL for executable checks. If the preference needs cross-file state, changed-file path ownership, human review, or richer semantics than a stable GritQL pattern can provide, keep it as
status: draft.
- Do not delete, disable, or weaken rules to make a task pass unless the user explicitly asks for that.
- Before finishing code changes, run
harness-lint check --changed and report any diagnostics.
- Use
harness-lint rule list after changing rule configuration. It prints Markdown tables; do not pass --json to rule list.
Project Sources
AGENTS.md is the source for durable repository workflow and coding constraints.harness.toml configures local rules and external packs.harness.lock records installed pack revisions..harness/ is generated cache and must stay uncommitted.
