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 RockyHong/super-bootstrap --skill triagegit clone --depth 1 https://github.com/RockyHong/super-bootstrapWrote 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/rockyhong/super-bootstrap/triage)<a href="https://agentmods.dev/skills/rockyhong/super-bootstrap/triage"><img src="https://agentmods.dev/badge/skills/rockyhong/super-bootstrap/triage.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 | $0.00159 | $0.01230 |
| Opus 5 | $0.00079 | $0.00615 |
| Sonnet 5 | $0.00032 | $0.00246 |
| Haiku 4.5 | $0.00016 | $0.00123 |
Grade A, and why
triage 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 — 37 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Triage — Read-Only Grounding Phase
Grounding pickup lane for a card. The thinking runs in the triage subagent (agents/triage.md, model: opus); this skill is the dispatch shell + absorption protocol. The agent grounds the card cold — premise verify, aim validate, blast collect — sizes the fix, appends a Verdict block to the card; the fix is a later phase that inherits the verdict as its context scope.
Arguments
| Invocation | Behavior |
|---|---|
/super-bootstrap:triage BUG-012 (any BUG/DEBT/GAP-### ID) |
Dispatch the triage subagent on that card. |
/super-bootstrap:triage (bare) |
List open cards with no Verdict block yet; the user picks one, then dispatch. |
Execution
- Resolve the card: Glob
docs/work/first (loads no file content, so the check fires nodocs/**path-scoped rule) — nothing back → the repo carries no card substrate, whatever else tracks its work; print "No runway installed. Run/super-bootstrapto set up the pipeline." and stop. Substrate present → the card filedocs/work/{ID}.mdexists; missing → report "no card {ID}", stop. A Verdict block already present in the card → surface its kind and date instead of re-dispatching (re-triage only on explicit user ask). - Dispatch:
Agenttool,subagent_type: "triage", prompt = the card ID + today's date + the gateway-aligned problem-aim when framing sharpened or corrected the card claim (premise / problem / scenario only). Exclude cause theories and fix preferences (bias-input exclusion) — the aligned aim is the user-validated target, not a prior; the card row carries the frozen claim. - Absorb the agent's report (
agents/triage.md§ Reporting):- DONE / DONE_WITH_CONCERNS — relay verdict + card path. auto-fix Verdict block → post the route line off its
Execution:tag (inline / phased → implement within the envelope, skipping what the tag skips; full → cluster route per CLAUDE.md). surface Verdict block → climb before any hand-off:docs/decisions.mdclosed forks, then the specs / project docs, then the aligned aim (one option contradicts it → the other wins), then the proxy the verdict'ssettles by:names, run in its sized container — aninline readruns here; adispatched probeorphased buildis a route change, so the framing + route line re-fires (CLAUDE.md § Framing + Route) and the measurement dispatches or lands as a plan block (gateway doesn't build — CLAUDE.md § Dispatch) — a fork that resolves there lands as a## Design — {date}block naming the settling source, and the route line follows; only a fork inside business / risk / preference / private authority reaches the user as the verdict's## Decision needed, then wait for the ruling. User picks an option (the verdict's recommendation as-is counts) → the gateway appends a## Design — {date}block (shape, and how a wait ruling names its party:docs/work/README.md§ Thread contract, Design line); user drops the card → delete the card file (resolve). - NEEDS_CONTEXT — relay the named gaps; the answer returns as an
## Amendmentappend on the card — via/super-bootstrap:log(its amend path) or directly by the answering session. - NEEDS_GRANTS — grant the named tooling and re-dispatch; user round-trip only when the grant itself is user-owned (cost, consent).
- BLOCKED — premise wrong; relay the counter-diagnosis to the user.
- DONE / DONE_WITH_CONCERNS — relay verdict + card path. auto-fix Verdict block → post the route line off its
- The verdict artifact rides the session's normal envelope commit — no in-phase commit.
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 Changed d9c20b8949a6
- 7d ago First seen · 37 lines · 159 tokens per session scan A 2236b310367a
triage is a skill published in the GitHub repository RockyHong/super-bootstrap (3 stars, last pushed 2d ago), licensed MIT. It adds 159 tokens to every session and 1,230 once invoked, about $0.0008 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
ln-51-opportunity-evaluator
Evaluates new product opportunities using demand, channels, competition, and economics. Not for committed-backlog or implementation planning.
{{BUCKET}}-management
Skill "{{BUCKET}}-management" from damusix/atomic-claude, covering purpose, doc shape and frontmatter contract.
plan-to-tasks
Transform a plancreation planv{N}.md into a structured tasks.yaml + per-WP workpackages/.yaml using parallel Sonnet sub-agents. Use after plan-creation finishes when you need execution-friendly task structure.
dev-finish
Close out a session — summarise what changed, capture what was learned, and prepare the commit.
dev-cycle
Run a feature end to end — scout the code, plan it, build it, and keep a session file so a closed terminal doesn't lose the thread.
answer-open-question-with-alternative
Record a chosen alternative from an open question's embedded analysis as that question's answer in the current milestone's requirements.md.