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 skills add tomsej/pi-ext --skill wf-reviewgit clone --depth 1 https://github.com/tomsej/pi-extWrote 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/skills/tomsej/pi-ext/wf-review)<a href="https://agentmods.dev/skills/tomsej/pi-ext/wf-review"><img src="https://agentmods.dev/badge/skills/tomsej/pi-ext/wf-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/skills/tomsej/pi-ext/wf-review"><img src="https://agentmods.dev/badge/skills/tomsej/pi-ext/wf-review.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00948 |
| Opus 5.5 | $0.00015 | $0.00379 |
| Sonnet 5.5 | $0.00007 | $0.00190 |
| Haiku 4.5 | $0.00004 | $0.00095 |
Grade A, and why
wf-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.
How it starts
The opening of the file, as written. The whole thing — 79 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/wf-review
Argument: optional absolute path to a contract. Without a contract, review the current diff. You own the review: reviewers only report; verify every finding and delegate fixes sequentially. Never edit production code yourself.
GATE = node ~/Workspace/pi-ext/wf/wf-gate.mjs
1. Scope
TARGET = sc worktree status --json → target_branch (fallback main). Run
ocr delegate preview --from TARGET --to HEAD, restore excluded product files
such as .md when they carry the change or the contract names them, then load
rules with ocr delegate rule <files>. Keep lockfiles and generated code out.
Write .wf/review-packet.md: contract path or “no contract”, refs, merge_base,
final files, diff stats, and rules. Every reviewer gets it. Without ocr, use
git diff --stat TARGET...HEAD and disclose the fallback.
2. Pick the panel
With a contract, use its exact GATE agents <spec> --json → review plan;
never infer models from YAML. Without a contract, inspect the packet and use:
- docs/tests-only with no runtime behavior or sensitive config: 1 reviewer;
- behavior changes: 2 reviewers for correctness and tests/edge cases;
- security, auth, permissions, secrets, migrations, data integrity, concurrency, public APIs, or a large cross-module diff: 3 reviewers for correctness, the relevant risk, and impact/tests.
Read PI_PROVIDER and PI_MODEL. Reviewers must differ by model from the
current session and each other. Prefer claude/fable,
codex/gpt-5.6-sol, pi/zai/glm-5.2, then
pi/anthropic/claude-opus-5. Skip matching or unavailable models, record the
fallback, and never add reviewers merely because more models exist.
3. Review — max 2 rounds
Run each round's reviewers IN PARALLEL with subagent_spawn, each in a fresh
context; never collapse them onto one model. Give each the packet, focus,
read-only mode, and these rules:
- with claude harness: the prompt MUST start with
/code-review; codex would treat/reviewas plain text, so other harnesses get no slash prefix; - with pi harness: use
sem_impactfor blast radius/tests andsem_context; - findings are CLAIMS: severity,
file:line, problem, and impact; with a contract also cite the violated criterion or scope; - Critical/High = bug, security, data loss, or uncovered criterion; Medium = performance or missing error handling with context; discard Low unless useful;
- project rules are the checklist; read-only means never
gh pr createorgh pr merge— for the codex harness this sentence is the only guard.
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 · 79 lines · 37 tokens per session scan A 857b0d4bd4c7
wf-review is a skill published in the GitHub repository tomsej/pi-ext (79 stars, last pushed 1mo ago), licensed MIT. It adds 37 tokens to every session and 948 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 2026-08-30.
Other skills, from other repositories
autoreview
Structured code review through the shared OpenClaw agent-skills installation.
omh-code-review
This is an OMH code-review workflow skill, projected for Agent Skills hosts (Claude Code, Codex, Cursor, opencode, OpenClaw, pi).
revdiff-plan
Review the last Codex assistant message (plan, analysis, or proposal) with inline annotations in a TUI overlay. Extracts the most recent response from Codex rollout files and opens it in revdiff for review and annotation. Activates on "revdiff-plan", "review plan with revdiff", "annotate plan", "review last response"…
re0-merge
Review and land an external contribution the way this suite does: gate it against the thesis, land it with the author's credit intact, complete a new skill rather than merging it raw, then approve, credit, and explain before closing. Use when reviewing a pull request, as any collaborator or maintainer, not only the…
explore-unknowns
Guide the user through a quadrant walk that maps the unknowns of a task — open by listing the known knowns, then work through known unknowns, unknown knowns, and unknown unknowns one stage at a time, ending with a complete four-quadrant map in the user's hands. Use when a request is ambiguous or underspecified, the…
asplos-writing-style
Use when drafting or revising ASPLOS prose — engineering the first two pages that rapid review actually reads, stating the cross-layer insight as one quotable sentence, structuring per-layer mechanism sections, writing bounded claims, and fitting everything into 11 self-contained pages that include figures and…