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 XRenSiu/claude-code-forge --skill acceptance-fleetgit clone --depth 1 https://github.com/XRenSiu/claude-code-forgeWrote 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/xrensiu/claude-code-forge/acceptance-fleet)<a href="https://agentmods.dev/skills/xrensiu/claude-code-forge/acceptance-fleet"><img src="https://agentmods.dev/badge/skills/xrensiu/claude-code-forge/acceptance-fleet/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/xrensiu/claude-code-forge/acceptance-fleet"><img src="https://agentmods.dev/badge/skills/xrensiu/claude-code-forge/acceptance-fleet.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 | $0.00335 | $0.10666 |
| Opus 5.5 | $0.00134 | $0.04266 |
| Sonnet 5.5 | $0.00067 | $0.02133 |
| Haiku 4.5 | $0.00034 | $0.01067 |
Grade C, and why
acceptance-fleet scanned grade C with 1 finding 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.
Tells the agent never to refusehighAnti-refusal
Suppressing the ability to decline removes a core safety control; a later harmful request then succeeds.
- **Missing `spec-robustness.md`** — do NOT refuse; proceed with a warning that S2.5 was skipped and `/spec-gaming-detector` runs in max-suspicion mode (its own iron rule 10 activates). How it starts
The opening of the file, as written. The whole thing — 504 lines — stays where its author put it; the contents beside it link to each section on GitHub.
acceptance-fleet — Step 5-6 orchestrator (v1.0+ pure dispatch)
You are invoked to verify an implementation against a done_when.yaml contract by dispatching the 6 standalone review skills, synthesizing their findings via /meta-judge, and decoding the verdict into a four-state ratchet decision (DONE / FIX / SPEC_DRIFT / GAMING_RISK).
Say once at the start, then start working:
"I'm using the acceptance-fleet skill. I'll dispatch the 6 review skills in parallel against your impl, hand findings to /meta-judge, decode the verdict into the 4-state ratchet, and write a full ratchet-log/iteration-NNN/ trace."
Do not narrate further — just walk the phases.
What changed in v1.0.0 (read this once if migrating from v0.x)
| v0.x (monolithic fleet) | v1.0+ (orchestrator) |
|---|---|
| 7 evaluator roles embedded internally (test-runner / existence-checker / requirement-tracer / design-reviewer / adversarial-reviewer / edge-case-hunter / e2e-explorer) | 6 standalone skills dispatched: /code-reviewer × focuses, /qa-reviewer, /pm-reviewer, /spec-drift-detector, /spec-gaming-detector |
| Meta-judge was a fleet phase (S3) | /meta-judge is its own skill, invoked by this orchestrator |
references/role-prompts.md carried all 8 role prompts |
Each review skill owns its prompt + references; orchestrator just calls them |
references/finding-schema.yaml had all schemas |
Each review skill owns its own finding-schema.yaml; meta-judge owns the final-verdict schema |
gaming-detector-patterns.md, meta-judge-protocol.md lived in this skill |
Migrated to spec-gaming-detector and meta-judge respectively |
done_when.yaml had a fitness: layer |
Retired per HTML v2 §3.5; replaced with rules: layer consumed by /meta-judge |
If you have v0.x ratchet-log/ directories, they remain readable — the new orchestrator writes the same format under references/ratchet-log-format.md. The migration is at the skill definition level, not the artifact level.
What ships with it
40 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.
- assets/evaluators.yaml 1.9 KB
- eval/fixtures/done_when_bad_bands.yaml 272 B
- eval/fixtures/done_when_bands.yaml 289 B
- eval/fixtures/qa-measurements-leaky.yaml 323 B
- eval/fixtures/qa-measurements-metrics-low.yaml 312 B
- eval/fixtures/qa-measurements-metrics-partial.yaml 291 B
- eval/fixtures/qa-measurements-metrics.yaml 465 B
- eval/fixtures/qa-measurements-nested-leak.yaml 543 B
- eval/fixtures/qa-reviewer-full.yaml 2.6 KB
- eval/fixtures/ratchet-log/iteration-001/fleet-outputs/pm-reviewer.yaml 67 B
- eval/fixtures/ratchet-log/iteration-001/fleet-outputs/qa-reviewer.yaml 70 B
- eval/fixtures/ratchet-log/iteration-001/fleet-outputs/spec-gaming-detector.yaml 210 B
- eval/fixtures/ratchet-log/iteration-001/impl-snapshot.tar.gz 44 B
- eval/fixtures/ratchet-log/iteration-002/fleet-outputs/pm-reviewer.yaml 67 B
- eval/fixtures/ratchet-log/iteration-002/fleet-outputs/qa-reviewer.yaml 70 B
- eval/fixtures/ratchet-log/iteration-002/fleet-outputs/spec-gaming-detector.yaml 235 B
- eval/fixtures/ratchet-log/iteration-002/impl-snapshot.tar.gz 44 B
- eval/fixtures/ratchet-log/iteration-003/fleet-outputs/pm-reviewer.yaml 67 B
- eval/fixtures/ratchet-log/iteration-003/fleet-outputs/qa-reviewer.yaml 70 B
- eval/fixtures/ratchet-log/iteration-003/fleet-outputs/spec-gaming-detector.yaml 419 B
- eval/fixtures/ratchet-log/iteration-003/impl-snapshot.tar.gz 44 B
- eval/fixtures/review-complete/complete/qa-reviewer.yaml 285 B
- eval/fixtures/review-complete/complete/spec-gaming-detector.yaml 241 B
- eval/fixtures/review-complete/incomplete/qa-reviewer.yaml 285 B
- eval/fixtures/review-complete/incomplete/spec-gaming-detector.yaml 266 B
- eval/fixtures/review-complete/mismatch/qa-reviewer.yaml 336 B
- eval/fixtures/review-complete/mismatch/spec-gaming-detector.yaml 241 B
- eval/fixtures/review-complete/truncated/qa-reviewer.yaml 285 B
- eval/fixtures/review-complete/truncated/spec-gaming-detector.yaml 235 B
- eval/gate.json 1.6 KB
- eval/report.md 920 B
- references/evaluation-isolation-levels.md 7.4 KB
- references/fix-prompt-template.md 5.3 KB
- references/ratchet-log-format.md 12 KB
- references/skill-dispatch-matrix.md 12 KB
- scripts/meets_done_when.py 10 KB runs code
- scripts/next_iteration.py 11 KB runs code
- scripts/pick_evaluators.py 4.6 KB runs code
- scripts/qa_facts.py 6.5 KB runs code
- scripts/verify_review_complete.py 20 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.
- 1mo ago Changed · +13 lines 2947b462ee4b
- 1mo ago Changed · +174 lines de129563010b
- 2mo ago First seen · 317 lines · 335 tokens per session scan C a542c960c0d2
acceptance-fleet is a skill published in the GitHub repository XRenSiu/claude-code-forge (2 stars, last pushed 1mo ago), licensed MIT. It adds 335 tokens to every session and 10,666 once invoked, about $0.0013 per session on Opus 5.5. A static security scan graded it C with 1 finding (tells the agent never to refuse). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
adk-go-self-review
Review an ADK Go change the way a maintainer will — a fresh-context pass over the whole diff, five lenses (correctness and tests, scope, simplicity, style, adk-python parity), and the mutation check that proves your tests pin the change. Use before opening a PR, before any later push that changes code, and when asked…
adversarial-reviewer
Stress-test a code change for concrete correctness defects, unsafe assumptions, and failure modes. Use when the user explicitly asks for an adversarial review, a deep bug hunt, or to tear a change apart, or to check your own change before submitting it. Do not use for ordinary review, style feedback, or general…
go-testing
Trigger: Go tests, go test coverage, Bubbletea teatest, golden files. Apply focused Go testing patterns.
semgrep-rule-variant-creator
Creates language variants of existing Semgrep rules. Use when porting a Semgrep rule to specified target languages. Takes an existing rule and target languages as input, produces independent rule+test directories for each language.
brooks-sweep
Full-sweep mode: runs a unified analysis across all quality dimensions — code decay, architecture, tech debt, and test quality — then applies fixes directly to the codebase. Safe changes are auto-applied; risky changes are confirmed before execution. Drawing on twelve classic engineering books. Triggers when: user…
plugin-review
Review plugin quality with tiered checks and dependency scoping. Use for PR and pre-release audits.