shareAI-lab/learn-claude-code skill
agent harness engineering
Learn Claude Code teaches how to build agent harnesses—the operational environment and tooling that enable LLM-based agents to work effectively in specific domains. It covers the complete pattern of tool implementation, knowledge curation, context management, permissions, and task systems through a 20-lesson tutorial series, using Claude as the example model.
Learn the agent loop architecture and how to structure tool calls between model and harness Implement tools, knowledge systems, and context management for LLM agents in your domain Design permission boundaries, approval workflows, and execution isolation for agent operations Build task systems with dependency tracking and parallel execution across subagents Understand the distinction between model intelligence (trained) and harness engineering (coded) 33 CRITICAL✓ 12 HIGH✓ 93 MEDIUM
✓ CRITICAL/HIGH reflect AI-verified findings (false positives excluded) · MEDIUM/LOW/INFO are unverified scanner output
AI-verified (CRITICAL/HIGH): 45 likely real (100%) 7 false positive — excluded from CRITICAL/HIGH count above
Findings by checker · 2 high-signal, 2 mostly false-positive (hidden by default)
CHK-081 35 findings 33 likely 2 false positive
6% FP CHK-099 8 findings 8 likely
0% FP ▼ Show 2 checkers that are mostly false positives (9 findings) 138 findings click to expand
CHK-081 Command injection risk — exec/execSync with string interpolation in s06_subagent/code.py
s06_subagent/code.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in s19_mcp_plugin/code.py
s19_mcp_plugin/code.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in agents/s11_autonomous_agents.py
agents/s11_autonomous_agents.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in agents/s07_task_system.py
agents/s07_task_system.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in agents/s05_skill_loading.py
agents/s05_skill_loading.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in s18_worktree_isolation/code.py
s18_worktree_isolation/code.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in s07_skill_loading/code.py
s07_skill_loading/code.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in s03_permission/code.py
s03_permission/code.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in agents/s03_todo_write.py
agents/s03_todo_write.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in agents/s12_worktree_task_isolation.py
agents/s12_worktree_task_isolation.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in agents/s08_background_tasks.py
agents/s08_background_tasks.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in agents/s_full.py
agents/s_full.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in s04_hooks/code.py
s04_hooks/code.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in s02_tool_use/code.py
s02_tool_use/code.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in agents/s04_subagent.py
agents/s04_subagent.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in s01_agent_loop/code.py
s01_agent_loop/code.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in s05_todo_write/code.py
s05_todo_write/code.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in s11_error_recovery/code.py
s11_error_recovery/code.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in s20_comprehensive/code.py
s20_comprehensive/code.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in s10_system_prompt/code.py
s10_system_prompt/code.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in agents/s10_team_protocols.py
agents/s10_team_protocols.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in s16_team_protocols/code.py
s16_team_protocols/code.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in agents/s01_agent_loop.py
agents/s01_agent_loop.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in s17_autonomous_agents/code.py
s17_autonomous_agents/code.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in s09_memory/code.py
s09_memory/code.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in s14_cron_scheduler/code.py
s14_cron_scheduler/code.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in s12_task_system/code.py
s12_task_system/code.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in s13_background_tasks/code.py
s13_background_tasks/code.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in s15_agent_teams/code.py
s15_agent_teams/code.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in agents/s06_context_compact.py
agents/s06_context_compact.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in agents/s09_agent_teams.py
agents/s09_agent_teams.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in s08_context_compact/code.py
s08_context_compact/code.py
AI: likely real confirmed ▼
CHK-081 Command injection risk — exec/execSync with string interpolation in agents/s02_tool_use.py
agents/s02_tool_use.py
AI: likely real confirmed ▼
CHK-099 Potential IDOR — 'task_id' accessed without ownership check
s13_background_tasks/code.py
AI: likely real likely ▼
CHK-099 Potential IDOR — 'task_id' accessed without ownership check
s19_mcp_plugin/code.py
AI: likely real likely ▼
CHK-099 Potential IDOR — 'task_id' accessed without ownership check
s15_agent_teams/code.py
AI: likely real likely ▼
CHK-099 Potential IDOR — 'task_id' accessed without ownership check
s17_autonomous_agents/code.py
AI: likely real likely ▼
CHK-099 Potential IDOR — 'task_id' accessed without ownership check
s18_worktree_isolation/code.py
AI: likely real likely ▼
CHK-099 Potential IDOR — 'task_id' accessed without ownership check
s14_cron_scheduler/code.py
AI: likely real likely ▼
CHK-099 Potential IDOR — 'task_id' accessed without ownership check
agents/s_full.py
AI: likely real likely ▼
CHK-099 Potential IDOR — 'task_id' accessed without ownership check
agents/s08_background_tasks.py
AI: likely real likely ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in s19_mcp_plugin/code.py
s19_mcp_plugin/code.py
AI: likely real likely ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in s18_worktree_isolation/code.py
s18_worktree_isolation/code.py
AI: likely real likely ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in s20_comprehensive/code.py
s20_comprehensive/code.py
AI: likely real likely ▼
CHK-125 os.system() — arbitrary shell execution — no scope constraint in code-review/SKILL.md
skills/code-review/SKILL.md
AI: likely real likely ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in s14_cron_scheduler/code.py
s14_cron_scheduler/code.py
possible ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in agents/s_full.py
agents/s_full.py
possible ▼
CHK-125b sudo invocation — privilege escalation — no scope constraint in agents/s_full.py
agents/s_full.py
possible ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in agents/s03_todo_write.py
agents/s03_todo_write.py
possible ▼
CHK-125b sudo invocation — privilege escalation — no scope constraint in agents/s03_todo_write.py
agents/s03_todo_write.py
possible ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in agents/s06_context_compact.py
agents/s06_context_compact.py
possible ▼
CHK-125b sudo invocation — privilege escalation — no scope constraint in agents/s06_context_compact.py
agents/s06_context_compact.py
possible ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in agents/s08_background_tasks.py
agents/s08_background_tasks.py
possible ▼
CHK-125b sudo invocation — privilege escalation — no scope constraint in agents/s08_background_tasks.py
agents/s08_background_tasks.py
possible ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in agents/s01_agent_loop.py
agents/s01_agent_loop.py
possible ▼
CHK-125b sudo invocation — privilege escalation — no scope constraint in agents/s01_agent_loop.py
agents/s01_agent_loop.py
possible ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in agents/s04_subagent.py
agents/s04_subagent.py
possible ▼
CHK-125b sudo invocation — privilege escalation — no scope constraint in agents/s04_subagent.py
agents/s04_subagent.py
possible ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in agents/s02_tool_use.py
agents/s02_tool_use.py
possible ▼
CHK-125b sudo invocation — privilege escalation — no scope constraint in agents/s02_tool_use.py
agents/s02_tool_use.py
possible ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in agents/s07_task_system.py
agents/s07_task_system.py
possible ▼
CHK-125b sudo invocation — privilege escalation — no scope constraint in agents/s07_task_system.py
agents/s07_task_system.py
possible ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in agents/s09_agent_teams.py
agents/s09_agent_teams.py
possible ▼
CHK-125b sudo invocation — privilege escalation — no scope constraint in agents/s09_agent_teams.py
agents/s09_agent_teams.py
possible ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in agents/s05_skill_loading.py
agents/s05_skill_loading.py
possible ▼
CHK-125b sudo invocation — privilege escalation — no scope constraint in agents/s05_skill_loading.py
agents/s05_skill_loading.py
possible ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in agents/s10_team_protocols.py
agents/s10_team_protocols.py
possible ▼
CHK-125b sudo invocation — privilege escalation — no scope constraint in agents/s10_team_protocols.py
agents/s10_team_protocols.py
possible ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in agents/s11_autonomous_agents.py
agents/s11_autonomous_agents.py
possible ▼
CHK-125b sudo invocation — privilege escalation — no scope constraint in agents/s11_autonomous_agents.py
agents/s11_autonomous_agents.py
possible ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in agents/s12_worktree_task_isolation.py
agents/s12_worktree_task_isolation.py
possible ▼
CHK-125b sudo invocation — privilege escalation — no scope constraint in agents/s12_worktree_task_isolation.py
agents/s12_worktree_task_isolation.py
possible ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in s09_memory/README.en.md
s09_memory/README.en.md
possible ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in s09_memory/code.py
s09_memory/code.py
possible ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in s07_skill_loading/code.py
s07_skill_loading/code.py
possible ▼
CHK-125b sudo invocation — privilege escalation — no scope constraint in s07_skill_loading/code.py
s07_skill_loading/code.py
possible ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in s06_subagent/code.py
s06_subagent/code.py
possible ▼
CHK-125b sudo invocation — privilege escalation — no scope constraint in s06_subagent/code.py
s06_subagent/code.py
possible ▼
CHK-125a write_file() — unrestricted file write — no scope constraint in s16_team_protocols/README.en.md
s16_team_protocols/README.en.md
possible ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in s16_team_protocols/code.py
s16_team_protocols/code.py
possible ▼
CHK-125a write_file() — unrestricted file write — no scope constraint in s16_team_protocols/README.md
s16_team_protocols/README.md
possible ▼
CHK-125a write_file() — unrestricted file write — no scope constraint in s16_team_protocols/README.ja.md
s16_team_protocols/README.ja.md
possible ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in s04_hooks/code.py
s04_hooks/code.py
possible ▼
CHK-125b sudo invocation — privilege escalation — no scope constraint in s04_hooks/code.py
s04_hooks/code.py
possible ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in s08_context_compact/code.py
s08_context_compact/code.py
possible ▼
CHK-125b sudo invocation — privilege escalation — no scope constraint in s08_context_compact/code.py
s08_context_compact/code.py
possible ▼
CHK-125a write_file() — unrestricted file write — no scope constraint in s12_task_system/README.en.md
s12_task_system/README.en.md
possible ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in s12_task_system/code.py
s12_task_system/code.py
possible ▼
CHK-125a write_file() — unrestricted file write — no scope constraint in s12_task_system/README.md
s12_task_system/README.md
possible ▼
CHK-125a write_file() — unrestricted file write — no scope constraint in s12_task_system/README.ja.md
s12_task_system/README.ja.md
possible ▼
CHK-125b sudo invocation — privilege escalation — no scope constraint in s20_comprehensive/code.py
s20_comprehensive/code.py
possible ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in s02_tool_use/code.py
s02_tool_use/code.py
possible ▼
CHK-125b sudo invocation — privilege escalation — no scope constraint in s02_tool_use/code.py
s02_tool_use/code.py
possible ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in s03_permission/code.py
s03_permission/code.py
possible ▼
CHK-125b sudo invocation — privilege escalation — no scope constraint in s03_permission/code.py
s03_permission/code.py
possible ▼
CHK-125b sudo invocation — privilege escalation — no scope constraint in s03_permission/README.md
s03_permission/README.md
possible ▼
CHK-125b sudo invocation — privilege escalation — no scope constraint in s03_permission/README.ja.md
s03_permission/README.ja.md
possible ▼
CHK-125a write_file() — unrestricted file write — no scope constraint in s11_error_recovery/README.en.md
s11_error_recovery/README.en.md
possible ▼
CHK-125 subprocess module — arbitrary shell execution — no scope constraint in s11_error_recovery/code.py
s11_error_recovery/code.py
possible ▼
CHK-125a write_file() — unrestricted file write — no scope constraint in s11_error_recovery/README.md
s11_error_recovery/README.md
possible ▼
▼ Show 7 false positives (5% of this view) Last scanned: Jul 12, 2026
More servers
google-gemini/gemini-cli 85
An open-source AI agent that brings the power of Gemini directly into your terminal.
106k★
skypilot-org/skypilot 85
obra/superpowers 85
Foundational skill pack by Jesse Vincent now in anthropics/claude-plugins-official. Includes ffuf web-fuzzing/pentest skill. Partial analysis done — full hook and plugin inspection pending. tier=T2
191k★
Significant-Gravitas/AutoGPT 85
AutoGPT is the vision of accessible AI for everyone, to use and to build on. Our mission is to provide the tools, so that you can focus on what matters.
186k★
nanocoai/nanoclaw 85
A lightweight alternative to OpenClaw that runs in containers for security. Connects to WhatsApp, Telegram, Slack, Discord, Gmail and other messaging apps,, has memory, scheduled jobs, and runs directly on Anthropic's Agents SDK
30k★
google-gemini/gemini-cli 85
An open-source AI agent that brings the power of Gemini directly into your terminal.
106k★
Scan your entire org's MCP deployment
2,500+ repos pre-scored. 22% carry CRITICAL findings.