brenocq/capcom
Mission control for your Claude Code spend
ARCHETYPE
Orchestrator
A bench of specialised subagents. The main agent mostly delegates.
Copy this rig
# review before running: this installs third-party code
$ npx degit brenocq/capcom/.claude ./rig-capcom # inspect, then merge into .claude/ MCP servers are added to Claude Code at local scope; env vars are shown as YOUR_… placeholders — we never store values. Files are fetched with degit into a separate folder so you can review before merging.
$ curl -fsSL --create-dirs -o .claude/agents/analyst.md https://raw.githubusercontent.com/brenocq/capcom/main/.claude/agents/analyst.md $ curl -fsSL --create-dirs -o .claude/agents/builder.md https://raw.githubusercontent.com/brenocq/capcom/main/.claude/agents/builder.md $ curl -fsSL --create-dirs -o .claude/agents/fixer.md https://raw.githubusercontent.com/brenocq/capcom/main/.claude/agents/fixer.md $ curl -fsSL --create-dirs -o .claude/agents/reviewer.md https://raw.githubusercontent.com/brenocq/capcom/main/.claude/agents/reviewer.md $ curl -fsSL --create-dirs -o .claude/agents/scout.md https://raw.githubusercontent.com/brenocq/capcom/main/.claude/agents/scout.md
This rig commits no guardrails. Here is the community baseline instead — the deny/ask rules most often found across all 16.2k rigs:
{
"permissions": {
"deny": [
"Read(./.env)",
"Read(./.env.*)",
"Read(**/.env)",
"Read(~/.ssh/**)",
"Bash(git push --force:*)",
"Read(**/*.pem)",
"Read(.env)",
"Read(~/.aws/**)",
"Read(**/*.key)",
"Read(**/.env.*)",
"Bash(git push --force*)",
"Bash(rm -rf:*)",
"Bash(rm -rf *)",
"Bash(sudo *)",
"Bash(sudo:*)",
"Bash(rm -rf /)",
"Bash(git push -f:*)",
"Read(.env.*)",
"Bash(git reset --hard*)",
"Bash(rm -rf /*)"
],
"ask": [
"Bash(git push:*)",
"Bash(git push *)",
"Bash(gh pr merge *)",
"Bash(git commit:*)",
"Bash(rm *)",
"Bash(gh pr merge:*)",
"Bash(gh pr create *)",
"Bash(git reset --hard:*)",
"Bash(npm publish:*)",
"Bash(git rebase *)"
]
}
} Subagents (5)
| analyst model: opus | Runs capcom on the real logs and judges its output (do the totals match Claude Code's own cost records, are the spike explanations and setup warnings right and useful), on Opus at medium effort, withi |
| builder model: opus | Builds one feature or module of capcom in the worktree its task names, on Opus at high effort, within a 150k-<redacted> budget. Use for new code and for changes that touch a module's design. |
| fixer model: sonnet | Fixes one bug or makes one small, well-specified change in capcom, on Sonnet at medium effort, within an 80k-<redacted> budget. Use when the cause or the change is already known. |
| reviewer model: sonnet | Reviews one branch of capcom before it merges, on Sonnet at medium effort, read-only, within a 100k-token budget. Use after a builder reports, on its diff against main. |
| scout model: haiku | Answers one factual question about the code or the log formats (where is X, which records carry Y, how often does Z happen), on Haiku at low effort, read-only, within a 40k-token budget. Use instead o |
Similar rigs
Gildaciolopes/agentic-dev-stack
Smart vibe coding with Claude Code: multi-agent orchestration (Opus plans; Sonnet, Haiku, Gemini, Codex and free models execute), Spec Kit SDD, caveman + RTK + graphify token savings, sandboxed secrets, self-healing free models and a Hermes
Minimalist 1.2k tok · · Q10
brackenjoe/agent-infrastructure
A guide for AI agent set up. A manager, a cheaper builder for code, any agents for other repetitive tasks, and the governance to keep them in line.
Minimalist 1.4k tok · · Q18
aniteshdharam/claude-token-policy
A Claude Code plugin that cuts your token bill by shrinking context per call, not by picking a cheaper model: a policy skill, enforcing hooks, and a census script that measures your own transcripts.
Minimalist 247 tok · · Q0
mutaaf/outatime-88
OUTATIME 88 — neon time-travel arcade racer, built and groomed by autonomous agents. Live at mutaafaziz.com/outatime
Pragmatist 11.5k tok · · Q3