jimy-r/agent-workspace-architecture
Agent-ready knowledge architecture, run daily in Claude Code: turn a coding agent into a system you can hand work to and trust while you're away. 18 load-bearing patterns, one-page workspace map, interactive tour, guided learn track, teardo
stack: Python
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 jimy-r/agent-workspace-architecture/samples/.claude ./rig-agent-workspace-architecture # 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.
$ npx degit jimy-r/agent-workspace-architecture/samples/.claude/skills/audit-workthrough .claude/skills/audit-workthrough $ npx degit jimy-r/agent-workspace-architecture/samples/.claude/skills/goal-design .claude/skills/goal-design $ npx degit jimy-r/agent-workspace-architecture/samples/.claude/skills/orient .claude/skills/orient $ npx degit jimy-r/agent-workspace-architecture/samples/.claude/skills/review-queue .claude/skills/review-queue $ npx degit jimy-r/agent-workspace-architecture/samples/.claude/skills/role-pressure-test .claude/skills/role-pressure-test $ npx degit jimy-r/agent-workspace-architecture/samples/.claude/skills/systematic-debugging .claude/skills/systematic-debugging $ npx degit jimy-r/agent-workspace-architecture/samples/.claude/skills/tasks .claude/skills/tasks $ npx degit jimy-r/agent-workspace-architecture/samples/.claude/skills/terse-mode .claude/skills/terse-mode $ npx degit jimy-r/agent-workspace-architecture/samples/.claude/skills/verify-completion .claude/skills/verify-completion $ npx degit jimy-r/agent-workspace-architecture/samples/.claude/skills/wrap .claude/skills/wrap
$ curl -fsSL --create-dirs -o .claude/agents/audit-second-opinion.md https://raw.githubusercontent.com/jimy-r/agent-workspace-architecture/main/samples/.claude/agents/audit-second-opinion.md $ curl -fsSL --create-dirs -o .claude/agents/audit.md https://raw.githubusercontent.com/jimy-r/agent-workspace-architecture/main/samples/.claude/agents/audit.md $ curl -fsSL --create-dirs -o .claude/agents/heartbeat.md https://raw.githubusercontent.com/jimy-r/agent-workspace-architecture/main/samples/.claude/agents/heartbeat.md $ curl -fsSL --create-dirs -o .claude/agents/researcher.md https://raw.githubusercontent.com/jimy-r/agent-workspace-architecture/main/samples/.claude/agents/researcher.md $ curl -fsSL --create-dirs -o .claude/agents/example-security.md https://raw.githubusercontent.com/jimy-r/agent-workspace-architecture/main/samples/example-project/.claude/agents/example-security.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 *)"
]
}
} Skills (10)
audit-workthroughgoal-designorientreview-queuerole-pressure-testsystematic-debuggingtasksterse-modeverify-completionwrap
Subagents (5)
| audit-second-opinion model: fable | Route here when the user asks for "a second opinion on the audit", "an independent audit" or a blind-spot check the primary audit missed, and on the quarterly pass or after a major audit.md refactor. |
| audit model: fable | Route here when the user says "run the audit", "/audit", "audit the setup" or "upgrade audit", or when the weekly audit cadence fires. Sweeps configs, hooks, security, MCP, memory and every project ag |
| heartbeat model: claude-opus-4-6 | RETIRED sample, studied predecessor only. Ran hourly on a scheduled task (Windows Task Scheduler / cron) — never chat-invoked, no routing trigger applies. Superseded by an explicit-delegation queue (s |
| researcher model: opus | Route here when the user says "research this", "find sources on" or "fact-check this", or asks anything needing outside evidence - market scans, library comparisons, regulatory/tax research, literatur |
| example-security model: claude-sonnet-4-6 | Security auditor for the Example Project. Invoke for security reviews, vulnerability assessments, or hardening audits specific to this project's codebase, credentials, and infrastructure. |
Similar rigs
jimy-r/agent-workspace-starter
Minimal governed agent-workspace template for Claude Code: orient/wrap session discipline, task capture, lessons loop, and two safety hooks on day one
Pragmatist 815 tok · · Q23
sohaibdevv/Claude-Starter-Kit
A pre-built configuration that gives Claude Code persistent memory, security guardrails, and structured workflows. Clone, run setup.sh, and your next session picks up where the last one left off.
Pragmatist 435 tok · · Q0
mnemcik/consigliere
Your trusted advisor for managing knowledge, projects, and ideas — powered by AI
Automator 7.8k tok · · Q20
DrCatHicks/learning-opportunities
A Claude or Codex skill for deliberate skill development during AI-assisted coding
Minimalist 601 tok · · Q0