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 ForgeyClap/claude-forge --skill forge-councilgit clone --depth 1 https://github.com/ForgeyClap/claude-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/forgeyclap/claude-forge/forge-council)<a href="https://agentmods.dev/skills/forgeyclap/claude-forge/forge-council"><img src="https://agentmods.dev/badge/skills/forgeyclap/claude-forge/forge-council/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/forgeyclap/claude-forge/forge-council"><img src="https://agentmods.dev/badge/skills/forgeyclap/claude-forge/forge-council.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 | $0.00042 | $0.00695 |
| Opus 5.5 | $0.00017 | $0.00278 |
| Sonnet 5.5 | $0.00008 | $0.00139 |
| Haiku 4.5 | $0.00004 | $0.00069 |
Grade A, and why
forge-council 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 19d 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 — 50 lines — stays where its author put it; the contents beside it link to each section on GitHub.
forge-council — LLM Council (FULL protocol)
Use ONLY when councilTrigger() (forge-quality.cjs) says FULL: an explicit owner request for
pushback, or high impact + high uncertainty. Never by default, never on simple tasks — the
expected information gain must outweigh latency, cost and context load. LIGHT exists as an outcome
of the trigger but deliberately has NO protocol here yet: first a benchmark that proves its value
(anti-overengineering); until then LIGHT falls back to one targeted second opinion.
Protocol (FULL)
Step 0 — neutral framing. Write the decision question NEUTRALLY (no preferred direction, no
"we lean towards X"): context, options, criteria, hard limits. Compute context_hash =
sha256 of that framing text (forge-quality.cjs::sha256). The framing is a run artifact.
Step 1 — 5 fresh, independent advisors, in parallel. Dispatch via the Agent tool (real subagents, one message with 5 parallel calls; NO simulated agents). Each advisor gets ONLY the neutral framing — no visibility of the others. The five fixed lenses:
- Contrarian — attacks the most popular option; looks for what everyone is missing.
- First Principles — reduces to fundamentals; ignores convention.
- Expansionist — widens the option space; what if the question itself is too narrow?
- Outsider — looks at it as an outsider/end user would; jargon-free.
- Executor — judges feasibility, cost, risk, rollback.
Step 2 — anonymous peer review. Anonymize the five pieces of advice as A-E and send them back to each advisor (a second dispatch round): rank the other four and name the strongest critique of each. Anonymity prevents authority bias and self-preference.
Step 3 — chairman synthesis. The orchestrator (or a separate chairman dispatch) weighs the advice + peer rankings and writes the decision WITH an explicit minority report: dissenting opinions don't disappear, they're recorded alongside the reason.
Record. Fill in a CouncilDecisionRecord: council_id, context_hash, participants (role + runtime +
dispatch_id from the REAL Agent dispatches + response_ref), responses, quorum, verdict,
minority_report, status. Validate with validateCouncilRecord and persist via
node .claude/forge-bin/forge-quality.cjs council-save <run_id> <record.json> — an invalid
record is NOT written (the refusal with a reason is the honest result), and a council_id is
never overwritten.
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.
- 19d ago Changed · -8 tokens per session 4ba376ec27e2
- 22d ago First seen · 50 lines · 50 tokens per session scan A 0c16491b6412
forge-council is a skill published in the GitHub repository ForgeyClap/claude-forge (2 stars, last pushed 16d ago), licensed MIT. It adds 42 tokens to every session and 695 once invoked, about $0.0002 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 2026-09-24.
Other skills, from other repositories
returns-policy
Answer return, refund, and warranty questions for electronics. Use when the user mentions returns, refunds, RMAs, warranty coverage, damaged items, or opened packaging.
technical-troubleshooting
Provide setup, troubleshooting, and maintenance guidance. Use when the user reports a device that won't power on, connectivity issues, setup questions, overheating, or maintenance concerns.
contribute-spec-fix
Fix a broken OpenAPI spec in jentic-public-apis with an OpenAPI Overlay, validate it (spectral lint + idempotency check), and contribute it back via a PR to the community catalog. Falls back to applying the same overlay to the local Jentic registry if the user can't wait for maintainer approval, and closes the loop by…
jentic
Calls third-party APIs through a Jentic One instance — governed, credential-managed access where the agent searches for an operation, inspects it, and executes it through an audited broker that holds the credentials (identity → discover → check access → execute). Applies when the user asks to call an external API or…
review
Use when operating as the Fabrik Review stage agent. This skill guides code review of an implementation, finding and fixing issues, and ensuring the PR is ready for human review.
implement-comment
Use when operating as the Fabrik Implement comment reviewer. This skill guides applying user-requested code changes during implementation, committing and pushing, and updating the task checklist — without signaling stage completion.