systematic-debugging

Warn

Audited by Gen Agent Trust Hub on Apr 15, 2026

Risk Level: MEDIUMDATA_EXFILTRATIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [DATA_EXFILTRATION]: The skill provides instructions and bash examples for diagnostic instrumentation that access sensitive information. This includes commands to list system keychains (security list-keychains), verify identities (security find-identity), and inspect environment variables (env | grep IDENTITY). These are presented as necessary steps for debugging multi-component system failures and environment configuration issues.
  • [COMMAND_EXECUTION]: The skill includes a utility script find-polluter.sh that executes tests using npm test to identify state pollution. It also provides various shell command examples in SKILL.md for environmental verification, such as codesign and security tools.
  • [PROMPT_INJECTION]: The skill employs strong behavioral constraints ('The Iron Law', 'ALWAYS', 'NEVER') to enforce a specific debugging persona. It also contains several test files (test-pressure-1.md, etc.) that simulate high-pressure scenarios (production outages, social pressure) designed to evaluate the agent's resistance to bypassing defined safety and procedural protocols.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) as its primary function involves ingesting and analyzing untrusted data such as error messages, stack traces, and system logs.
  • Ingestion points: Error messages, stack traces, and logs are actively processed by the agent during Phase 1 (Root Cause Investigation) as specified in SKILL.md and root-cause-tracing.md.
  • Boundary markers: Absent. The instructions do not define delimiters or provide warnings for the agent to ignore instructions embedded within the logs or error messages being analyzed.
  • Capability inventory: The skill possesses capabilities to execute shell commands via example snippets and the find-polluter.sh script.
  • Sanitization: Absent. There is no instruction for the agent to sanitize or escape the content of error messages or logs before processing them or using them to inform subsequent command execution.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Apr 15, 2026, 12:15 AM
Security Audit — agent-trust-hub — systematic-debugging