Security●●Solid
Agent-pd – A zero-token audit log to catch rogue Claude Code subagents
Catches denied calls that never reach transcripts—six detectors at zero token cost.
Big BrainNiche Gem
softie123
625d ago
Back to browse
GitHub Repository
A police department for your Claude Code agents — a logging-only hook + CLI that audits the main agent and every subagent and reports rule offenses (permission bypass, out-of-scope & credential access, self-permissioning, disallowed tools, redundant, off-task) with quoted evidence. Catch-and-report, never blocks.
16 starsPython
A police department for your Claude Code agents
by softie123·Jun 11, 2026·10 points·8 comments
AI Analysis
●●SolidWizardryNiche Gem
Hash-chained audit log catches agent violations at zero token cost.
Strengths
- •Hash-chained tamper-evident audit log with optional off-host append-only sink
- •Six deterministic detectors catch specific violations without LLM token costs
- •Covers main agent plus all subagents including Workflow tool spawns
Weaknesses
- •Logging-only approach catches violations but doesn't prevent them
- •Claude Code-specific, won't work with other agent frameworks
Category
Target Audience
Developers building with Claude Code agents
Similar To
MLflow · Weights & Biases
Similar Projects
AI/ML●●●Banger
Multicorn Shield – Open-source permissions and approvals for AI agents
MCP proxy intercepts agent tool calls; no code changes needed. Actual safeguard for runaway agents.
Zero to OneSolve My ProblemBig Brain
rachelle-r
123mo ago
Security●●●Banger
I accidentally caught an AI agent trying to poison my prod config
Hash-chained audit logs caught Claude Code attempting to inject staging endpoints into production config.
Big BrainWizardrySolve My Problem
zippolyon
103mo ago
Security●●●Banger
ScopeGate – Granular permission gateway for AI agents (MCP, open-core)
MCP permission proxy solves real AI agent over-permissioning—88% of orgs hit this problem.
Solve My ProblemBig BrainZero to One
jetbootsmaker
103mo ago
Security●●●Banger
Merkle Mountain Range audit log and execution tickets for AI agents
Merkle Mountain Range ledger proves AI agents can't retroactively fake logs—novel crypto primitive.
Big BrainZero to OneWizardry
nknutalapati
313mo ago
AI/ML●●Solid
Axon – Agentic AI with mandatory user approval and audit logging
Agent approval gates and audit logs beat open-source alternatives, but multi-agent governance isn't novel.
Solve My ProblemBig Brain
NeuroVexon
123mo ago