skf-rename-skill

Pass

Audited by Gen Agent Trust Hub on May 6, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill performs significant file system operations including recursive copying, directory moving, and recursive deletion (rm -rf) of old skill directories. These operations are core to the skill's primary purpose of renaming assets and are implemented with a transactional safety workflow that includes copy-before-delete logic and mandatory verification steps.
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it ingests and processes content from external files like .export-manifest.json and various skill metadata files to determine file paths and content updates.
  • Ingestion points: Reads .export-manifest.json in step-01-select.md and reads SKILL.md, metadata.json, context-snippet.md, and provenance-map.json in step-02-execute.md.
  • Boundary markers: The skill explicitly uses <!-- SKF:BEGIN --> and <!-- SKF:END --> delimiters in platform context files to scope its modifications.
  • Capability inventory: Performs file moves, recursive deletions, and content writes across the local filesystem in step-02-execute.md.
  • Sanitization: Implements JSON parsing, schema version validation (schema_version: "2"), and structural verification before proceeding to irreversible deletion operations.
Audit Metadata
Risk Level
SAFE
Analyzed
May 6, 2026, 03:12 AM