android-anr-investigator
Pass
Audited by Gen Agent Trust Hub on Aug 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from ANR traces, stack dumps, and logs to identify potential issues and uses extracted values as arguments for shell commands. This creates an indirect prompt injection attack surface.\n
- Ingestion points: ANR traces, main thread stacks, and logs in SKILL.md.\n
- Boundary markers: Present; the skill suggests wrapping suspected blocking classes or methods in double quotes in the codegraph command template.\n
- Capability inventory: Shell execution of codegraph explore, rg, and find as specified in the CodeGraph Integration section.\n
- Sanitization: Absent; the skill does not specify explicit validation or filtering for data extracted from logs before use in commands.\n- [COMMAND_EXECUTION]: The skill provides instructions for executing shell-based tools (rg, find, and codegraph) to explore the codebase and identify blocking calls.
Audit Metadata