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/asteasolutions/ai-toolkit/helmnpx skills add asteasolutions/ai-toolkit --skill helmgit clone --depth 1 https://github.com/asteasolutions/ai-toolkitWhat 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 | $0.00044 | $0.01598 |
| Opus 5 | $0.00022 | $0.00799 |
| Sonnet 5 | $0.00009 | $0.00320 |
| Haiku 4.5 | $0.00004 | $0.00160 |
Grade A, and why
helm 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 2d 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 — 88 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Helm
The developer is at the helm. You sequence and gate; they decide. Conduct by invoking the phase skills — never play their instruments yourself.
Invariant: helm sequences and gates — it never decides. When unsure whether something is a decision, treat it as one and gate it.
Invariant: each phase runs by invoking its skill. If you write a spec, grill, cut tickets, or compose a reconcile delta yourself, you skipped an invocation. Inline a phase only when the Skill tool reports it unavailable — and tell the developer.
Invariant: helm dispatches the legwork — it never does it. Reading source, searching the tree, editing code, running tests: each belongs to a sub-agent. If you opened a source file yourself, you skipped a dispatch. The work tree under
.scratch/<goal-slug>/is the only thing you read and write directly.
Dispatch
Legwork runs in a sub-agent so its intermediate tokens never enter this conversation — the developer can still open the sub-agent to watch it think. Use the harness's sub-agent tool (Task in Claude Code, runSubagent in Copilot, Agent in Cursor). Do the legwork inline only when no sub-agent tool exists in this harness — and say so before you start.
Prefer the repo's specialist to a generic sub-agent. Once per run, before the first dispatch, take stock of what is available: the agent types the sub-agent tool offers, plus any agent definitions the repo keeps (commonly .claude/agents/). Match each subtask to the narrowest agent whose brief covers it — a read-only explorer to orient, a named specialist wherever one fits the work — and fall back to general-purpose only when none does. Name the agent you chose at the gate, so the developer can correct the match.
A dispatch prompt is self-contained — the sub-agent shares none of your context and cannot ask you a question. Paste the leaf doc's relevant sections into the prompt; never refer to "the plan above". Name the phase skill the sub-agent must run.
What ships with it
3 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.
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.
- 2d ago First seen · 88 lines · 44 tokens per session scan A 209b40128b98
helm is a skill published in the GitHub repository asteasolutions/ai-toolkit (5 stars, last pushed 5d ago), licensed MIT. It adds 44 tokens to every session and 1,598 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 2026-08-31.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
brainstorming
You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.
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…
agent-host-chat-contributions
Build and review cross-cutting agent-host chat behavior through lifecycle contributions. Use when adding turn lifecycle side effects, prompt or context injection, restored-history transformation, protocol-action observation, or when reviewing changes that add code to AgentSideEffects or AgentService.
auto-perf-optimize
Run agent-driven VS Code performance or memory investigations. Use when asked to launch Code OSS, automate a VS Code scenario, run the Chat memory smoke runner, capture renderer heap snapshots, take workflow screenshots, compare run summaries, or drive a repeatable scenario before heap-snapshot analysis.