Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add commands/burin-labs/harn/pr-finish-passgit clone --depth 1 https://github.com/burin-labs/harnWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00000 | $0.02955 |
| Opus 5 | $0.00000 | $0.01477 |
| Sonnet 5 | $0.00000 | $0.00591 |
| Haiku 4.5 | $0.00000 | $0.00296 |
Grade A, and why
pr-finish-pass scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 279 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Run a final-polish pass on the current PR or feature branch before it lands.
The goal is the state a senior reviewer would land without comments. The sweep is targeted at the recurring issues this repo accumulates between "the feature works" and "the PR is landable" — not a generic code review.
Phase 0 — identify the work
git status --short
git rev-parse --abbrev-ref HEAD
git log --oneline origin/main..HEAD
gh pr view --json number,state,headRefName,mergeStateStatus 2>/dev/null
If gh pr view returns a PR, you are in PR mode. Otherwise branch mode —
skip rebase/CI phases and go straight to the review sweep.
Always read the full diff before touching anything: git diff origin/main...HEAD.
In PR mode, also read existing review comments — sometimes the easiest finding
is one a reviewer already left.
Phase 1 — rebase on origin/main (PR mode only)
git fetch origin main
git rebase origin/main
Conflict rules in this repo:
CHANGELOG.md: keep both sides; preserve the top heading.docs/theme/harn-keywords.js: regenerate withmake gen-highlight.docs/src/language-spec.md: regenerate via the pre-commit hook (editspec/HARN_SPEC.mdinstead).Cargo.lock: resolve from the merged manifests, then runmake setup-rust.- Any file where both sides reformatted: take one side then run the
appropriate formatter (
make fmt,make fmt-harn, portal lint).
Then git push --force-with-lease. Never bare --force.
Phase 2 — drive CI to green (PR mode only)
gh pr checks
gh run list --branch "$(git branch --show-current)" --limit 5
gh run view <run-id> --log-failed # for each failure
Address every failure comprehensively — no continue-on-error papering, no
silenced asserts. Common categories:
- Format / clippy / lint:
make fmt,make lint,make lint-harn,make fmt-harn. Commit the fixes. - Test failure: reproduce locally with
make test-oneand theHARN_TEST_ONE_NAMEandHARN_TEST_ONE_PACKAGEvariables, then fix the root cause. Usemake test-affectedwhen the failure spans a changed crate. - Conformance:
./scripts/harn_bin.sh -- test conformance --filter <name>. If.expectedneeds an update for an intentional behavior change, do it explicitly. - Portal / VS Code / tree-sitter: run the relevant
npm run portal:lint|build|test/(cd tree-sitter-harn && npm test). - Drift checks (
check-language-spec,check-highlight,check-trigger-quickref): re-run the generators and commit the outputs. make lint-test-patterns: see Phase 3.2.check-docs-snippets: a.harnsnippet in docs no longer parses; fix the snippet, not the parser.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 279 lines · 0 tokens per session scan A 4db05803dbb1
pr-finish-pass is a command published in the GitHub repository burin-labs/harn (20 stars, last pushed yesterday), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 2,955 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other commands, from other repositories
callee-create-agent
Create a Callee agent or deterministic workflow.
callee
Route a task through configured Callee agents and workflows.
delegate
Delegate to Cursor, Composer, or cursor-agent via cursor-delegate-mcp (see delegate skill).
design-review
Workflow recipe — review a design end-to-end, ending in measured numbers rather than adjectives, by chaining 4 skills.
setup-pm-skills
Onboard a new user — find out what they do, recommend the right bundles & top skills, and set up a project CONTEXT.md so every skill is tailored to them.
security-review
CWE 기반 보안 검토 + STRIDE 위협 모델링 (v6 - effort:max 강제).