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 kubev2v/forklift-console-plugin --skill reviewergit clone --depth 1 https://github.com/kubev2v/forklift-console-pluginWrote 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/kubev2v/forklift-console-plugin/reviewer)<a href="https://agentmods.dev/skills/kubev2v/forklift-console-plugin/reviewer"><img src="https://agentmods.dev/badge/skills/kubev2v/forklift-console-plugin/reviewer/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/kubev2v/forklift-console-plugin/reviewer"><img src="https://agentmods.dev/badge/skills/kubev2v/forklift-console-plugin/reviewer.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 | $0.00077 | $0.02429 |
| Opus 5.5 | $0.00031 | $0.00972 |
| Sonnet 5.5 | $0.00015 | $0.00486 |
| Haiku 4.5 | $0.00008 | $0.00243 |
Grade A, and why
reviewer 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 26d 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 — 203 lines — stays where its author put it; the contents beside it link to each section on GitHub.
PR Reviewer (orchestrator)
Output you see vs what gets posted
| Shown to you only | Posted to GitHub |
|---|---|
| Context loaded (rules-manifest.json) | — |
| Lessons consulted (lessons-context.md) | — |
| preview.md (event, full main body, inline table + links) | Same content as review.json after you confirm Post? |
Lessons and rule paths must not appear in review.json unless a finding stands on its own without citing the lesson.
Modes
| Mode | Triggers | After review |
|---|---|---|
| Peer review (default) | review PR, review #N, PR URL(s) |
Full preview → your edits → Post? → post-review.sh |
| Self review | self review, review my PRs |
Full preview → suggest fixing (no post). Fix phase → pr-monitor per PR |
Strictness
- Ask strictness before analysis unless the user stated it:
light|standard|strict|most strict - Self review defaults to
most strict - Peer default:
standard
| Level | Expectation |
|---|---|
light |
Blockers and clear bugs only |
standard |
Bugs, missing tests for new exports, a11y/i18n gaps, inconsistent patterns |
strict |
Edge cases, parallel flows, consumer blast radius, weak assertions |
most strict |
Above + consistency nits, docs, under-tested paths |
Multiple PRs
- Extract all PR numbers from input.
- One PR → orchestrator workflow below.
- Multiple PRs → one child orchestrator per PR in parallel (each runs steps 2–10); parent rolls up previews.
- Self-review fixes → one PR at a time; then pr-monitor (unchanged).
Orchestrator workflow (single PR)
You are the orchestrator. Do not substitute a single-agent review for the six perspective Tasks.
- Confirm mode + strictness.
.agents/skills/reviewer/scripts/gather-context.sh <PR>.agents/skills/reviewer/scripts/select-lessons-for-pr.sh <PR>.agents/skills/reviewer/scripts/select-rules-for-pr.sh <PR>- Launch six perspective subagents in parallel, using the corresponding Codex custom agent. Each subagent must:
- Follow the matching prompt in
.agents/skills/reviewer/prompts/(developer.md,qe.md,ux.md,architect.md,security.md,forklift-expert.md). - Pass
<PR>, strictness, and paths fromrules-manifest.jsonfor that perspective. - Write output to
/tmp/pr-review-<PR>/perspectives/<slug>.json(developer,qe,ux,architect,security,forklift-expert).
- Follow the matching prompt in
- After all Tasks complete, run
.agents/skills/reviewer/scripts/merge-perspective-reviews.sh <PR>
Or merge manually intoreview.jsonif subagents need correction (same schema as below). .agents/skills/reviewer/scripts/validate-review.sh <PR>— fix and re-validate until pass..agents/skills/reviewer/scripts/build-review-preview.sh <PR>- Present to the user in order (not a recap-only summary):
- Context loaded — from
rules-manifest.json(shared+ each perspective). - Lessons consulted — contents of
lessons-context.md. - GitHub draft — paste or summarize
preview.md: event, full main comment text, inline table (path, line, GitHub link, comment, side).
- Context loaded — from
- Peer: If the user edits findings, update
review.json, re-validate, rebuild preview. On confirm Post? →post-review.sh <PR>.
Self review: Do not ask to post; offer fix workflow.
What ships with it
15 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- agents/openai.yaml 43 B
- prompts/architect.md 578 B
- prompts/developer.md 1.1 KB
- prompts/forklift-expert.md 693 B
- prompts/qe.md 651 B
- prompts/security.md 535 B
- prompts/ux.md 540 B
- scripts/build-review-preview.sh 2.5 KB runs code
- scripts/gather-context.sh 3.9 KB runs code
- scripts/merge-perspective-reviews.sh 2.2 KB runs code
- scripts/merge-posted-review.sh 2.3 KB runs code
- scripts/post-review.sh 603 B runs code
- scripts/select-lessons-for-pr.sh 3.6 KB runs code
- scripts/select-rules-for-pr.sh 4.6 KB runs code
- scripts/validate-review.sh 5.6 KB runs code
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.
- 26d ago First seen · 203 lines · 77 tokens per session scan A d2c7edb509d6
reviewer is a skill published in the GitHub repository kubev2v/forklift-console-plugin (11 stars, last pushed today), licensed Apache-2.0. It adds 77 tokens to every session and 2,429 once invoked, about $0.0003 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 2026-10-21.
reviewer compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| reviewerkubev2v/forklift-console-plugin | Skill | 11 | today | Apache-2.0 | A | 77 | Claude Code, Codex |
| autoreviewopenclaw/clawhub | Skill | 9.6k | today | MIT | A | 17 | Claude Code, Codex |
| agent-teams-simplify-and-hardenLeoYeAI/openclaw-master-skills | Skill | 2.2k | 3d ago | MIT | A | 115 | Claude Code, Codex |
| omh-code-reviewrlaope/oh-my-hermes | Skill | 5.7k | today | MIT | A | 20 | Claude Code |
| revdiff-planumputun/revdiff | Skill | 957 | today | MIT | A | 84 | Claude Code |
| re0-mergeLilMGenius/paperthin | Skill | 1.4k | today | MIT | A | 51 | Claude Code, Codex |
| explore-unknownsdzhng/skills | Skill | 1.5k | 2d ago | MIT | A | 162 | Claude Code, Codex |