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 agentmods add skills/generous-corp/pulp/forge-modularnpx skills add Generous-Corp/pulp --skill forge-modulargit clone --depth 1 https://github.com/Generous-Corp/pulpWrote 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/generous-corp/pulp/forge-modular)<a href="https://agentmods.dev/skills/generous-corp/pulp/forge-modular"><img src="https://agentmods.dev/badge/skills/generous-corp/pulp/forge-modular.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 | $0.00028 | $0.29034 |
| Opus 5 | $0.00014 | $0.14517 |
| Sonnet 5 | $0.00006 | $0.05807 |
| Haiku 4.5 | $0.00003 | $0.02903 |
Grade A, and why
forge-modular 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 today.
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 — 2,123 lines — stays where its author put it; the contents beside it link to each section on GitHub.
forge-modular — generating VCV Rack modules and patches, and proving it
Covers tools/rack/ (the generator, the patch idiom library, the gates),
examples/forge-modular/ (the module pack), and forge-seam/ (the app's
sources, which live here and are built inside a Forge worktree).
Read this before trusting any green result in this area. Everything below cost a session at least once, and most of it was a test or a tool reporting success for work it had not done.
A run that fails still has to hand something over
generate() returns (patch, why, shortfall). shortfall is None on a pass
and a Shortfall when the loop ran out of attempts and is handing over the
best thing it built. It only raises SystemExit when there is genuinely
nothing to hand over — no attempt ever got past the lint.
The rule this encodes: an advisory verdict is not fatal, and out of attempts
is not the same as nothing to show. The loop used to raise SystemExit("gave up after 5 attempts") after generating five patches its own transcript had
called "structurally sound" and "keeping the patch anyway", and the user got
nothing — not the patches, not a copyable reason. "Not a sequenced-voice patch
yet" is a wording that means the loop should keep going, not that the patch
is worthless.
Consequences to keep in step when touching this:
- The build writes the handover as
<slug>-unfinished.vcv, prints the "built N modules, M cables → path" line the app parses for an artifact, then thehandover_report()block, and exits 1. Dropping either the exit code or thegave up afterphrase breaks a different reader: the shell decides a run ended by classifying that phrase, so without it the app watches a dead build forever. - A patch that failed the lint never becomes a
Shortfall. It names modules Rack cannot create, so handing it over offers a file that will not open. Shortfall.rankis(severity, misses, -attempt), lowest wins: a patch that plays beats one measured silent, fewer missed requirements beats more, and only a tie goes to the later attempt. Rank on requirements missed, never on the number of LINES it takes to say so — naming installed jacks adds lines, and ranking on those would rate a patch worse for every jack we managed to name.keep_attemptis not gated onFORGE_ATTEMPT_DIRany more. It was, and nothing butprove_idioms.shever set it, so every app build kept nothing while the log claimed otherwise. The default is a stamped per-pid directory under~/Library/Application Support/Forge Modular/attempts/.
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.
- today First seen · 2,123 lines · 28 tokens per session scan A 9967ad08e059
forge-modular is a skill published in the GitHub repository Generous-Corp/pulp (16 stars, last pushed today), licensed MIT. It adds 28 tokens to every session and 29,034 once invoked, about $0.0001 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 2026-09-04.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
local-ai-agents
Build local-first AI agents that run entirely on a developer workstation with Microsoft Foundry Local and Qwen function-calling models. Covers Small Language Models (SLMs), the OpenAI-compatible local endpoint, sandboxed local tools, local RAG with Chroma, local MCP servers, hybrid cloud/local routing, and the…
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
chat-pet-sprite-creation
Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…
insight-error-page
Write or audit an insight-kind error page for the Next.js dev overlay. Use when creating a new errors/ .mdx page, auditing an existing one, or checking that a page matches the framework fix cards. Covers page structure, title alignment, FixCard cards with Copy prompt button, code snippets, terminology verification…