skills-index-updater
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a bundled Python script (
update_skill_index.py) to perform its indexing tasks. This script handles file scanning and index file updates. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted metadata from
SKILL.mdfiles to generate indexes, which represents a potential injection surface. - Ingestion points: Reads metadata from
SKILL.mdfiles in global (e.g.,~/Documents/Cline/skills/) and local (e.g.,.claude/skills/) directories. - Boundary markers: The extracted content is placed within designated sections such as
## Available Global Skills Indexand## Available Local Skills Index. - Capability inventory: The associated script performs file reads and writes (
Path.read_text,Path.write_text) on documentation files. It does not perform network operations or execute content from the scanned files. - Sanitization: The script uses
yaml.safe_loadfor parsing frontmatter to prevent deserialization attacks. It does not perform HTML or Markdown escaping on descriptions, though the impact is limited to formatting in the generated index files.
Audit Metadata