Borrowing it
Nothing to install: this file belongs to ForgeyClap/claude-forge. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/ForgeyClap/claude-forge/main/.claude/skills/forge-worktrees/SKILL.mdgit 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-worktrees)<a href="https://agentmods.dev/skills/forgeyclap/claude-forge/forge-worktrees"><img src="https://agentmods.dev/badge/skills/forgeyclap/claude-forge/forge-worktrees.svg" alt="Measured on agentmods" 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 | $0.00837 |
| Opus 5 | $0.00021 | $0.00418 |
| Sonnet 5 | $0.00008 | $0.00167 |
| Haiku 4.5 | $0.00004 | $0.00084 |
Grade A, and why
forge-worktrees 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 4d 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 — 57 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Forge playbook — Safe parallel work via git worktrees
Do not duplicate ECC skills — defer to: using-git-worktrees (detect existing isolation → native
tools → git worktree fallback, mechanics and commands). This file is the Forge-specific orchestration
wrapper: when Forge reaches for isolation, and who owns merging the result back.
Hard rules
- Never let two agents write concurrently to the same working tree. If Head Chef dispatches parallel
Bosses on the same repo, each gets its own worktree/branch — no exceptions, per the project's
parallel-vs-serial doctrine (
CLAUDE.md→ "Parallel vs serial"). - Detect existing isolation before creating anything new (per
using-git-worktreesStep 0) — don't nest worktrees or fight a harness that already manages workspace lifecycle. - Only isolate genuinely independent subtasks (reference
dispatching-parallel-agents). Dependent steps stay serial in the main tree — a worktree does not make a dependency parallel-safe. - Never force-push, reset --hard, or otherwise destructively touch the integration branch without explicit owner approval, even when resolving a merge conflict.
- A non-source hotspot (event registries, sync manifests, dashboard files, doctor/config) that doesn't
warrant a full worktree still needs the
forge-lock-guard.cjs acquire/releaseprotocol around its dispatch (seeforge-routerStep 4) — lock-guard and worktree isolation are complementary safeguards, not either/or.
When to isolate
- Head Chef is running two or more Bosses concurrently against the same repo (e.g. Build Boss + UI Boss on different modules of the same app in the same run).
- A large/risky change (major refactor, dependency upgrade) that should stay reviewable as its own diff before merging into the main line.
- Codex write-delegation (
/codex:rescueor any--writetask) — perCODEX_GLOBAL_POLICY.md, Codex runs on a separate worktree/branch, never concurrently with Claude in the same tree.
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.
- 4d ago First seen · 57 lines · 42 tokens per session scan A b27d8ad9e644
forge-worktrees is a skill published in the GitHub repository ForgeyClap/claude-forge (2 stars, last pushed 1mo ago), licensed MIT. It adds 42 tokens to every session and 837 once invoked, about $0.0002 per session on Opus 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 2026-09-03.
Other skills, from other repositories
flame-harness-submit
Phase 10 — upload store text metadata + categories via fastlane, then pause with exact manual steps for the final iOS review submission and Android production promotion.
flame-harness-build
Phase 8 — bootstrap signing credentials, generate fastlane config from templates, and build + upload signed IPA (TestFlight) and AAB (internal track).
fabrik-implement
Use when operating as the Fabrik Implement stage agent. This skill guides the implementation of a planned feature, following the task checklist to produce committed, tested, pushed code on a feature branch.
fabrik-review-comment
Use when operating as the Fabrik Review comment reviewer. This skill guides applying user decisions on review findings — fixing issues, dismissing false positives, or deferring items — then committing and pushing without signaling stage completion.
workspace-init
Scan a multi-repo workspace root (a directory whose children are independent git repos) and generate a workspace navigation harness: a catalog CLAUDE.md router, per-repo .claude/repo-index/ .md deep indexes read on demand, .claude/meta/navigation.md, and a .claude/harness/seams/ cluster config ready for refresh-seams.…
pr-comments
Triage unresolved review threads on an open PR and close them out. Per thread, either agree (fix scoped to THIS PR's diff, verify, commit and push, reply, resolve) or disagree (justify with a quoted citation from the version-matched official doc, reply, resolve). Every reply is drafted for approval before it posts.…