debug-investigate

Pass

Audited by Gen Agent Trust Hub on Apr 29, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses shell commands like grep and glob to search the codebase. It also executes a local Python script (.claude/scripts/code_graph) to query structural information like callers, importers, and code traces. These operations are performed locally and are consistent with a developer's debugging workflow.
  • [DATA_EXFILTRATION]: The skill writes investigation findings to local directories such as plans/reports/ and .ai/workspace/analysis/. There are no instructions or patterns indicating the transmission of data to external servers or non-whitelisted domains.
  • [PROMPT_INJECTION]: The skill includes numerous guardrails to ensure agent reliability, such as 'Red Flag Stop Conditions' that mandate escalation to the user via AskUserQuestion for security-sensitive code or architectural changes. It does not contain instructions to ignore system prompts or safety filters.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 29, 2026, 09:18 AM
Security Audit — agent-trust-hub — debug-investigate