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.
git clone --depth 1 https://github.com/Pierry/harness-kitWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/pierry/harness-kit/review)<a href="https://agentmods.dev/commands/pierry/harness-kit/review"><img src="https://agentmods.dev/badge/commands/pierry/harness-kit/review/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/pierry/harness-kit/review"><img src="https://agentmods.dev/badge/commands/pierry/harness-kit/review.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What 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 5.1 | $0.00037 | $0.00510 |
| Opus 5.5 | $0.00015 | $0.00204 |
| Sonnet 5.5 | $0.00007 | $0.00102 |
| Haiku 4.5 | $0.00004 | $0.00051 |
Grade A, and why
review 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 1mo ago.
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.
What it actually says
Review a System Design Doc. Follow .claude/agents/system-architect/guides/pipeline.md, and .claude/shared/pipeline-pattern.md for inputs (resolve-mark-proceed) and eval (adversarial).
Resolve the design-doc path, do not ask (resolve-mark-proceed): use the path the user passes; if none
is given, default to the latest under .claude/runtime/outputs/architect/design/. Reviews own or
external designs. Never stop to ask for it.
Read:
- the target design doc
- .claude/agents/system-architect/guides/design-method.md
- .claude/agents/system-architect/guides/writing-style.md
- .claude/agents/system-architect/guides/templates/design-review.md
- .claude/agents/system-architect/guides/pipeline.md
Interrogate, do not summarize. Use the skill .claude/agents/system-architect/skills/review/SKILL.md. Default skeptical. Verdict: ship | revise | block.
Save to .claude/runtime/outputs/architect/review/{feature_id}.md (reuse design feature_id).
Sensors: .claude/agents/system-architect/sensors/design-structure.md (review variant). Evals: .claude/agents/system-architect/evals/design-review-depth.md.
Run the evals adversarially: dispatch a fresh evaluator via the Task tool (subagent_type: general-purpose) that did not author this review. Hand it only the artifact path and the one rubric path; it scores against the rubric and reports the weighted total plus the low-scoring dimensions. Below threshold (8.0) retries per pipeline.md, regenerating only the flagged dimensions.
After save, reply with this exact shape:
Review saved at {path}. Verdict: {ship|revise|block}. Score: {N}/10.
sensors: design-structure ok (10 questions, verdict present)
eval: design-review-depth {N}/10 (attempts: N)
refs: design/{feature_id}.md
blockers: {list or none}
next: revise the design, or proceed to /product-manager:prp or /sse:plan
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.
- 1mo ago First seen · 38 lines · 37 tokens per session scan A ca3b67e36db3
review is a command published in the GitHub repository Pierry/harness-kit (4 stars, last pushed 12d ago), licensed MIT. It adds 37 tokens to every session and 510 once invoked, about $0.0001 per session on Opus 5.5. 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 2026-08-31.
Other commands, from other repositories
assemble-team
Assemble a pre-built agent team for parallel work - review, feature, debug, cross-platform, full-stack, or research.
review-code
Trigger a comprehensive code review on recent changes or specified files.
clean-check
Analyze code for cleanliness issues (unused code, comment quality, formatting, naming, complexity). Delegates to the code-cleanliness agent.
review-branch
Review an existing branch holistically before merging — blast radius, conventions, security, and spec compliance.
code-review
Review staged git changes for bugs, security issues, and style violations.
p5-review-sprint
Runs ONE final code review over the entire sprint at sprint end, with a stronger model, to catch issues that single-story reviews structurally cannot see: cross-story interactions, the schema/data model as a whole, conformance to ADRs and Constitution Inviolables, and consistency with prior lessons/instincts.