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.
git clone --depth 1 https://github.com/JotJunior/cstkWrote 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/commands/jotjunior/cstk/agente-00c)<a href="https://agentmods.dev/commands/jotjunior/cstk/agente-00c"><img src="https://agentmods.dev/badge/commands/jotjunior/cstk/agente-00c.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 | $0.00056 | $0.18446 |
| Opus 5 | $0.00028 | $0.09223 |
| Sonnet 5 | $0.00011 | $0.03689 |
| Haiku 4.5 | $0.00006 | $0.01845 |
Grade D, and why
agente-00c scanned grade D with 3 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.
Asks for rootmediumPrivilege escalation
A mod that escalates privileges can change anything on the machine, not only the project.
> no descritor (`chmod 600`) e no prompt do spawn (SEC-H3: roteamento por Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
escreve em `<projeto-alvo>/.claude/settings.json`, que pode estar Reads MCP configurationmediumAgent snooping
mcp.json carries server URLs and auth tokens; reading it lets a mod discover and abuse other integrations.
if [ -f "<PAP>/.mcp.json" ] && grep -q '"cstk-state"' "<PAP>/.mcp.json" 2>/dev/null; then How it starts
The opening of the file, as written. The whole thing — 1,242 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/agente-00c
Voce vai iniciar uma nova execucao do orquestrador autonomo agente-00C
conforme contrato em docs/specs/_archived/agente-00c/contracts/cli-invocation.md.
Fronteira command↔orquestrador (lock + init): este command PAI detem o lock (acquire no passo 3, release SEMPRE no passo 5.ter) e inicializa o
state.json. O orquestrador NAO adquire lock nem re-inicializa estado — contrato canonico em "Fronteira command↔orquestrador" deagente-00c-orchestrator.md. Identico no resume (/agente-00c-resume).
Argumentos recebidos
$ARGUMENTS
Comportamento esperado
0. Warm-up de permissoes (CRITICO — antes de qualquer outra coisa)
A pipeline 00C invoca dezenas de skills/tools ao longo de horas/ondas. Permissoes pedidas "lazy" (no momento de cada invocacao) quebram a autonomia se o operador nao estiver presente para responder o prompt no instante exato — fluxo trava aguardando humano.
Solucao: invocar TODAS as skills/tools que serao usadas em batch no inicio, ANTES do parse de args. Cada invocacao dispara o prompt de permissao UMA vez aqui, com o operador ainda na sessao. Apos o warm-up, o orquestrador roda autonomamente sem interrupcoes.
Apresente ao operador:
Agente-00C — Warm-up de permissoes
Vou agora invocar cada skill/tool que sera usada na pipeline para
disparar TODOS os prompts de permissao em batch. Voce sera questionado
sobre cada uma; aprove para autorizar a execucao autonoma posterior.
Apos esse batch, o agente roda sem mais interrupcoes — suas respostas
de permissao aqui valem para todas as ondas subsequentes.
Continuar? [s/N]
Se o operador confirmar, execute em sequencia (cada item dispara o prompt nativo do Claude Code uma vez):
| # | Tool/Skill | Modo de warm-up |
|---|---|---|
| 1 | tool Skill — briefing |
invocar com --help ou prompt minimo "responda apenas OK" |
| 2 | tool Skill — constitution |
idem |
| 3 | tool Skill — specify |
idem |
| 4 | tool Skill — clarify |
idem |
| 5 | tool Skill — plan |
idem |
| 6 | tool Skill — checklist |
idem |
| 7 | tool Skill — create-tasks |
idem |
| 8 | tool Skill — execute-task |
idem |
| 9 | tool Skill — review-task |
idem |
| 10 | tool Skill — review-features |
idem |
| 11 | tool Agent — agente-00c-orchestrator |
spawn com prompt "warm-up: responda apenas READY" + subagent_type correto |
| 12 | tool Agent — agente-00c-clarify-asker |
idem |
| 13 | tool Agent — agente-00c-clarify-answerer |
idem |
| 14 | tool ScheduleWakeup | invocacao com delaySeconds: 60 + prompt: "warm-up no-op" + reason: "agente-00C warm-up" (depois cancele se possivel via re-schedule muito longo, ou aceite o no-op de 60s) |
| 15 | tool Bash — script state-rw.sh |
--help (no-op) |
| 16 | tool Bash — script bash-guard.sh check-blocklist --command "true" |
exit 0 sem efeito |
| 17 | tool Bash — git --version |
dispara permissao para git |
| 18 | tool Bash — gh --version |
dispara permissao para gh |
| 19 | tool Read — qualquer arquivo do projeto-alvo (ex: <PAP>/.git/HEAD se existir, senao <PAP> listing) |
dispara permissao para Read |
| 20 | tool Write — <PAP>/.claude/agente-00c-warmup-test (depois delete) |
dispara permissao para Write em <PAP> |
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 Changed · +17 lines b3cf83f808db
- 8d ago First seen · 1,225 lines · 56 tokens per session scan D 1f845271bd75
agente-00c is a command published in the GitHub repository JotJunior/cstk (23 stars, last pushed 4d ago), licensed MIT. It adds 56 tokens to every session and 18,446 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it D with 3 findings (asks for root, reads agent configuration directories, reads mcp configuration). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other commands, from other repositories
checklist
Generate a custom checklist for the current feature based on user requirements.
clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.
specify
Create or update the feature specification from a natural language feature description.
converge
Assess the current codebase against the feature's spec, plan, and tasks, then append any remaining unbuilt work as new tasks to tasks.md so implement can complete it.
implement
Execute the implementation plan by processing and executing all tasks defined in tasks.md.
analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.