optimize-agent-docs

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • PROMPT_INJECTION (LOW): The skill is susceptible to indirect prompt injection because it processes untrusted documentation as a primary input.\n
  • Ingestion points: The skill explicitly inventories and reads content from project-specific files like CLAUDE.md, AGENTS.md, and directories such as .cursor/ and .claude/.\n
  • Boundary markers: No explicit delimiters or system-level instructions are provided to ensure the agent ignores executable instructions found within the documentation files it is processing.\n
  • Capability inventory: The agent is instructed to extract facts and 'Critical Rules' and write them into a KNOWLEDGE.md manifest, which could allow a malicious instruction in a source file to be promoted to a high-priority agent rule.\n
  • Sanitization: There is no evidence of sanitization or content validation applied to the data before it is re-integrated into the agent's optimized knowledge base.\n- COMMAND_EXECUTION (SAFE): The skill utilizes standard, read-only shell utilities for file discovery.\n
  • Evidence: Workflow utilizes find (with a limited -maxdepth 2), grep, and wc to analyze and validate documentation structure.这些命令是非破坏性的且局限于本地工作区。
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:31 PM