lisa-wiki-ingest

Pass

Audited by Gen Agent Trust Hub on Aug 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest data from external and potentially untrusted sources, which could contain malicious instructions designed to influence the agent's synthesis or repository actions.
  • Ingestion points: SKILL.md specifies the ingestion of "GitHub, Linear, document, or transcript knowledge."
  • Boundary markers: The instructions do not specify the use of delimiters or provide warnings to the agent to ignore instructions embedded within the ingested data.
  • Capability inventory: The skill grants the agent the ability to execute git commit, git push, and utilize the gh CLI to create and auto-merge pull requests targeting the main branch.
  • Sanitization: Although the skill instructs the agent not to preserve secrets or credentials, it lacks sanitization or escaping protocols for natural language instructions found in the source documents.
  • [COMMAND_EXECUTION]: The skill requires the agent to execute shell commands for repository maintenance and metadata retrieval.
  • Evidence: The workflow involves running "verification checks" and using the gh (GitHub CLI) tool to fetch PR metadata and manage the pull request lifecycle.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 17, 2026, 06:27 AM
Security Audit — agent-trust-hub — lisa-wiki-ingest