skill-content-postmortem
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from user-provided content and JSON metrics, creating a surface where malicious instructions could potentially be embedded.\n
- Ingestion points: The skill accepts '内容原文或链接' (Content text or link) and '多条内容数据' (Multiple content data records) as primary inputs.\n
- Capability inventory: The agent is instructed to execute a local Python script 'scripts/aggregate.py' and write resulting reports to the 'outputs/' directory.\n
- Boundary markers: There are no instructions providing delimiters or 'ignore' commands for the content being analyzed.\n
- Sanitization: The instructions do not define validation or sanitization steps for the ingested external content.\n- [COMMAND_EXECUTION]: The skill instructions command the agent to execute shell commands to aggregate data.\n
- The agent is directed to run 'python3 skills/builtin/skill-content-postmortem/scripts/aggregate.py' using user-provided JSON files as input.\n- [DYNAMIC_EXECUTION]: The skill relies on the runtime execution of a local Python script to perform its core analytical functions.\n
- The execution of 'scripts/aggregate.py' is a dynamic process triggered by user input, which is a standard functional component but constitutes a dynamic execution pattern.
Audit Metadata