publish-skill

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill acts on untrusted data by reading and modifying files from a user-specified source directory. This creates a surface where malicious instructions embedded in those files could be executed by the agent.
  • Ingestion points: SKILL.md and all files within the <source_skill_path> directory.
  • Boundary markers: None; the agent processes the file contents directly as text for auditing and generalization.
  • Capability inventory: The skill uses Read, Write, Edit, Bash, Grep, and Glob tools across the SKILL.md and the scripts/audit_skill.sh file.
  • Sanitization: No specific escaping or instruction-blocking delimiters are used for the content being audited.
  • [COMMAND_EXECUTION]: The skill uses the Bash tool to execute a bundled audit script (scripts/audit_skill.sh). This script performs automated searches for sensitive information within a user-provided directory.
  • [DYNAMIC_EXECUTION]: The workflow relies on executing a local shell script to handle complex PII detection logic, including the optional use of exiftool for binary file metadata.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 01:59 AM
Security Audit — agent-trust-hub — publish-skill