routing-table-updater
Pass
Audited by Gen Agent Trust Hub on Aug 31, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses a pipeline of Python scripts to automate repository maintenance tasks, including scanning for skill files, extracting trigger phrases, and updating the central routing index (
commands/do.md). - [INDIRECT_PROMPT_INJECTION]: The skill processes metadata (descriptions and triggers) from external skill and agent files to generate routing tables. This creates a surface for indirect prompt injection if those source files contain malicious instructions, though the skill implements validation gates to mitigate accidental errors.
- [DYNAMIC_EXECUTION]: The skill instructions utilize an inline Python script via shell heredoc to validate the integrity of generated indices by checking for phantom file paths during the update phase.
- [DATA_EXPOSURE]: The skill scans the repository's
skills/andagents/directories to read metadata and YAML frontmatter for the purpose of building discovery indices.
Audit Metadata