~ / rigs / dralgorhythm / claude-agentic-framework

dralgorhythm/claude-agentic-framework

A More Effective Agent Harness for Claude

↗ GitHub ★ 125 no license updated 3mo ago project Claude Code Codex Claude plugin

share on X

ARCHETYPE

Fort Knox

Deny lists, pre-tool hooks, sandboxing. Nothing touches prod without a signature.

CONTEXT TAX · EVERY TURN

~17.3k tokens

Context hog · median rig: 2.3k · breakdown

GUARDRAILS

4/5

Blocks destructive commands · Protects secrets · Pre-tool screening hook · No YOLO mode · details

Copy this rig

# review before running: this installs third-party code
$ claude mcp add sequential-thinking -- npx -y @modelcontextprotocol/server-sequential-thinking
$ claude mcp add chrome-devtools -- npx -y chrome-devtools-mcp
$ claude mcp add filesystem -- npx -y @modelcontextprotocol/server-filesystem .
$ claude mcp add context7 -- npx -y @upstash/context7-mcp
$ npx degit dralgorhythm/claude-agentic-framework/.claude ./rig-claude-agentic-framework  # 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.

Merge into .claude/settings.json (project) or ~/.claude/settings.json (user). Hook commands reference scripts in the source repo — copy those too.

{
  "permissions": {
    "deny": [
      "// === Destructive File Operations ===",
      "Bash(rm -rf:*)",
      "Bash(rm -r:*)",
      "Bash(sudo:*)",
      "// === Git Destructive Operations ===",
      "Bash(git push --force:*)",
      "Bash(git push -f:*)",
      "Bash(git push origin --force:*)",
      "Bash(git push origin -f:*)",
      "Bash(git reset --hard:*)",
      "Bash(git clean -f:*)",
      "Bash(git clean -fd:*)",
      "// === Infrastructure Destructive Operations ===",
      "Bash(terraform destroy:*)",
      "Bash(terraform apply -auto-approve:*)",
      "Bash(docker system prune:*)",
      "Bash(docker rm:*)",
      "Bash(docker rmi:*)",
      "// === Dangerous Permission Changes ===",
      "Bash(chmod 777:*)",
      "// === Pipe-to-Shell (Remote Code Execution) ===",
      "Bash(curl * | bash*)",
      "Bash(wget * | bash*)",
      "Bash(curl * | sh*)",
      "Bash(wget * | sh*)",
      "// === Sensitive Credential Paths (Read) ===",
      "Read(~/.ssh/**)",
      "Read(~/.gnupg/**)",
      "Read(~/.aws/**)",
      "Read(~/.azure/**)",
      "Read(~/.config/gcloud/**)",
      "// === Shell Profile & SSH Key Protection (Edit/Write) ===",
      "Edit(~/.bashrc)",
      "Edit(~/.zshrc)",
      "Edit(~/.bash_profile)",
      "Write(~/.bashrc)",
      "Write(~/.zshrc)",
      "Write(~/.bash_profile)",
      "Edit(~/.ssh/**)",
      "Write(~/.ssh/**)",
      "// === Project Secrets (Read) ===",
      "Read(**/.env)",
      "Read(**/.env.*)",
      "Read(**/secrets/**)",
      "Read(**/*.pem)",
      "Read(**/id_rsa*)",
      "Read(**/credentials)",
      "Read(**/.npmrc)",
      "Read(~/.git-credentials)",
      "Read(~/.config/gh/**)",
      "// === Push to Default Branch ===",
      "Bash(git push origin main:*)",
      "Bash(git push origin master:*)",
      "Bash(git push -u origin main:*)",
      "Bash(git push -u origin master:*)"
    ]
  },
  "hooks": {
    "PreToolUse": [
      {
        "matcher": "Write|Edit|MultiEdit|NotebookEdit",
        "hooks": [
          {
            "type": "command",
            "command": "$CLAUDE_PROJECT_DIR/.claude/hooks/pre-tool-use-validator.sh"
          }
        ]
      },
      {
        "matcher": "Bash",
        "hooks": [
          {
            "type": "command",
            "command": "$CLAUDE_PROJECT_DIR/.claude/hooks/dangerous-command-guard.sh"
          },
          {
            "type": "command",
            "command": "$CLAUDE_PROJECT_DIR/.claude/hooks/pre-tool-use-validator.sh"
          },
          {
            "type": "command",
            "command": "$CLAUDE_PROJECT_DIR/.claude/hooks/pre-commit-verification.sh"
          },
          {
            "type": "command",
            "command": "$CLAUDE_PROJECT_DIR/.claude/hooks/pre-push-main-blocker.sh"
          },
          {
            "type": "command",
            "command": "$CLAUDE_PROJECT_DIR/.claude/hooks/branch-pr-discipline.sh"
          }
        ]
      }
    ]
  }
}

MCP servers (4)

server	source	est. tokens
Sequential Thinking · "sequential-thinking"	npm	900
Chrome DevTools MCP · "chrome-devtools"	npm	9.0k
Filesystem	npm	4.2k
Context7	npm	1.2k

Skills (12)

architect builder code-check land-the-plane qa-engineer security-auditor swarm-execute swarm-plan swarm-research swarm-review tailor ui-ux-designer

Subagents (5)

worker-architect model: opus	Senior architecture decisions. Use for complex design problems requiring deep analysis.
worker-builder model: sonnet	Implementation, testing, and refactoring worker for swarm tasks. Use for parallel coding, test writing, and code cleanup.
worker-explorer model: haiku	Codebase exploration and web research worker. Use for pattern search, dependency mapping, quick API/doc lookup, and library comparison.
worker-research model: sonnet	Deep research and investigation worker. Use for multi-source analysis, technology evaluation, competitive research, and comprehensive documentation.
worker-reviewer model: sonnet	Code review, security audit, and QA worker for swarm tasks. Use for parallel review, vulnerability detection, and quality assessment.

Hooks (12)

event	matcher	runs
SessionStart	startup\|resume\|compact\|clear	$CLAUDE_PROJECT_DIR/.claude/hooks/session-start-loader.sh
PreToolUse	Write\|Edit\|MultiEdit\|NotebookEdit	$CLAUDE_PROJECT_DIR/.claude/hooks/pre-tool-use-validator.sh
PreToolUse	Bash	$CLAUDE_PROJECT_DIR/.claude/hooks/dangerous-command-guard.sh
PreToolUse	Bash	$CLAUDE_PROJECT_DIR/.claude/hooks/pre-tool-use-validator.sh
PreToolUse	Bash	$CLAUDE_PROJECT_DIR/.claude/hooks/pre-commit-verification.sh
PreToolUse	Bash	$CLAUDE_PROJECT_DIR/.claude/hooks/pre-push-main-blocker.sh
PreToolUse	Bash	$CLAUDE_PROJECT_DIR/.claude/hooks/branch-pr-discipline.sh
PostToolUse	Edit\|MultiEdit\|Write	$CLAUDE_PROJECT_DIR/.claude/hooks/post-tool-use-tracker.sh
PostToolUse	Edit\|MultiEdit\|Write	$CLAUDE_PROJECT_DIR/.claude/hooks/post-edit-lint.sh
Stop	*	$CLAUDE_PROJECT_DIR/.claude/hooks/stop-validator.sh
SubagentStop	*	$CLAUDE_PROJECT_DIR/.claude/hooks/subagent-stop-validator.sh
TaskCompleted	*	$CLAUDE_PROJECT_DIR/.claude/hooks/task-quality-gate.sh

Permissions

deny (55)

// === Destructive File Operations ===

Bash(rm -rf:*)

Bash(rm -r:*)

Bash(sudo:*)

// === Git Destructive Operations ===

Bash(git push --force:*)

Bash(git push -f:*)

Bash(git push origin --force:*)

Bash(git push origin -f:*)

Bash(git reset --hard:*)

Bash(git clean -f:*)

Bash(git clean -fd:*)

// === Infrastructure Destructive Operations ===

Bash(terraform destroy:*)

Bash(terraform apply -auto-approve:*)

Bash(docker system prune:*)

Bash(docker rm:*)

Bash(docker rmi:*)

// === Dangerous Permission Changes ===

Bash(chmod 777:*)

// === Pipe-to-Shell (Remote Code Execution) ===

Bash(curl * | bash*)

Bash(wget * | bash*)

Bash(curl * | sh*)

Bash(wget * | sh*)

// === Sensitive Credential Paths (Read) ===

Read(~/.ssh/**)

Read(~/.gnupg/**)

Read(~/.aws/**)

Read(~/.azure/**)

Read(~/.config/gcloud/**)

// === Shell Profile & SSH Key Protection (Edit/Write) ===

Edit(~/.bashrc)

Edit(~/.zshrc)

Edit(~/.bash_profile)

Write(~/.bashrc)

Write(~/.zshrc)

Write(~/.bash_profile)

Edit(~/.ssh/**)

Write(~/.ssh/**)

// === Project Secrets (Read) ===

Read(**/.env)

Read(**/.env.*)

Read(**/secrets/**)

Read(**/*.pem)

Read(**/id_rsa*)

Read(**/credentials)

Read(**/.npmrc)

Read(~/.git-credentials)

Read(~/.config/gh/**)

// === Push to Default Branch ===

Bash(git push origin main:*)

Bash(git push origin master:*)

Bash(git push -u origin main:*)

Bash(git push -u origin master:*)

ask (0)

—

allow (120)

// === Core Claude Tools (Read/Navigation) ===

Read

Glob

Grep

WebSearch

// === Core Claude Tools (Write) ===

Edit

Write

NotebookEdit

// === File System Read Operations ===

Bash(cat:*)

Bash(ls:*)

Bash(head:*)

Bash(tail:*)

Bash(wc:*)

Bash(file:*)

Bash(stat:*)

Bash(diff:*)

Bash(tree:*)

Bash(find:*)

Bash(which:*)

Bash(type:*)

Bash(pwd:*)

Bash(xxd:*)

Bash(realpath:*)

Bash(dirname:*)

Bash(basename:*)

// === File System Write Operations ===

Bash(cp:*)

Bash(mv:*)

Bash(touch:*)

Bash(mkdir:*)

Bash(chmod:*)

// === Text Processing ===

Bash(grep:*)

Bash(sort:*)

Bash(uniq:*)

Bash(awk:*)

Bash(sed:*)

Bash(tr:*)

Bash(cut:*)

Bash(xargs:*)

Bash(jq:*)

Bash(yq:*)

// === System Info ===

Bash(date:*)

Bash(uname:*)

Bash(whoami:*)

Bash(id:*)

Bash(ps:*)

Bash(lsof:*)

Bash(env:*)

Bash(printenv:*)

Bash(netstat:*)

Bash(ss:*)

Bash(dig:*)

Bash(whois:*)

Bash(curl:*)

Bash(command -v:*)

Bash(command:*)

Context tax

Instructions: 1.2kMCP tool schemas (est.): 15.3kSkill metadata: 553Subagent metadata: 212

ALWAYS-LOADED FILES (claude-code)

≡

CLAUDE.md · 107 lines · ~1.2k tok

≡

AGENTS.md · 63 lines · ~1.1k tok

Estimate: characters ÷ 4 for instruction files and skill/subagent descriptions; MCP schema sizes from typical tool counts. Harnesses with deferred tool loading pay less. Method

Guardrails 4/5

✓

Blocks destructive commands

19 deny/ask rule(s) e.g. Bash(rm -rf:*)

✓

Protects secrets

Denies reads like // === Sensitive Credential Paths (Read) ===

✓

Pre-tool screening hook

6 PreToolUse hook(s)

✓

No YOLO mode

Permission prompts stay on

✗

Sandbox or ask-first rules

No sandbox and no ask-first rules

Badge

Add to your README:

[![AgentRigs](https://agentrigs.dev/badge/rig/dralgorhythm/claude-agentic-framework.svg)](https://agentrigs.dev/r/dralgorhythm/claude-agentic-framework)

Indexed from commit 6436a20fadad on 2026-10-09. Data is structure only — env values and anything secret-looking are stripped.