ClawAudit verdict

agent-dev-toolkit

agent-dev-toolkit-cahdieng

45
🟠 Risky
Significant concerns — only install if you understand the risks

The bundled 'Agent Wallet' component enables autonomous EVM-chain cryptocurrency transactions, real DeFi protocol interactions, and multi-chain asset management without explicit per-transaction user approval, representing elevated financial risk beyond typical dev tooling.

⚠ Flagged for review — coarse, uncorroborated signal, not a confirmed exploit. Review the config yourself before installing.

Automated static analysis — not a human review. ClawAudit flags capabilities, not confirmed intent, and can produce false positives. Disagree with this verdict? Use Dispute below.

10
security
90
transparency
90
maintenance

Permission integrity

Installs packages at runtime — transitive dependencies are not auditable

package_install

Findings (8)

Pattern match critical

Prompt injection — tries to override agent instructions

dependencies/agent-docs/references/advanced-patterns.md · code · Ignore previous instructions

Pattern match medium

References agent memory files

dependencies/agent-builder/references/templates.md · code · MEMORY.md

Pattern match medium

Accesses OpenClaw config/secrets directly

dependencies/agent-builder/references/openclaw-workspace.md · prose · downgraded · ~/.openclaw/openclaw.json

Pattern match medium

Accesses shell history/config

dependencies/agent-development/rules/agent-memory-limits.md · prose · downgraded · ~/.zshrc

Pattern match medium

Instructs covert action — may act without user awareness

dependencies/agent-development/rules/custom-agent-instructions.md · prose · downgraded · silently

Pattern match medium

Accesses sensitive environment variables

dependencies/agent-docs/references/advanced-patterns.md · code · process.env.API_KEY

Pattern match low

Makes HTTP request to external URL

dependencies/agent-development/rules/agent-pattern.md · code · fetch('https://

Pattern match low

References agent configuration files

dependencies/agent-docs/SKILL.md · prose · downgraded · CLAUDE.md

Why the tier is capped

Execution sink present in raw bytes (Hard Floor: class A/B). Final tier capped at Caution — cannot be lifted by any downgrade, example-payload opt-in, or allowlist.

Permissions & capabilities

Requires 1 system binary.

package_install

Is this flag fair?

Check another skill Browse the registry Auditing your own skills or configs? Use the API