Skills
Navigate
skills/jarmen423/skills/why/Gen Agent Trust Hub

why

Pass

Audited by Gen Agent Trust Hub on Sep 1, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data from multiple sources including Slack, Notion, Linear, and PR comments, creating a vulnerability to indirect prompt injection.
  • Ingestion points: MCP-integrated services (Slack, Notion, Linear, Sentry, Datadog, Databricks) and Git/GH output.
  • Boundary markers: Prompts in references/investigator-prompt.md and references/synthesizer-prompt.md enforce a strict investigator posture and citation-based verification.
  • Capability inventory: Spawning subagents, executing shell commands (git, gh), and SQL read operations.
  • Sanitization: Procedural sanitization via epistemic discipline and mandatory spot-checking of citations.
  • [COMMAND_EXECUTION]: The skill uses git and gh (GitHub CLI) to retrieve repository history and pull request metadata.
  • [DYNAMIC_EXECUTION]: The skill orchestrates parallel investigation and synthesis by spawning specialized subagents using the generalPurpose type.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 1, 2026, 02:19 AM