p-ego-search
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external platforms which creates a vulnerability surface.
- Ingestion points: Slack messages, Fireflies transcripts, and GitHub issues/PRs/discussions (SKILL.md).
- Boundary markers: The skill provides instructions for the agent to paraphrase content and avoid interpreting it as instructions, though no formal delimiters are established.
- Capability inventory: The skill executes shell commands via the
ghCLI and performs network operations through Slack and Fireflies MCP servers (SKILL.md). - Sanitization: Instructions recommend paraphrasing as a defense against direct echoing of potentially malicious instructions.
- [COMMAND_EXECUTION]: The skill utilizes the
gh(GitHub) CLI to perform searches across issues, pull requests, and discussions, and uses the GitHub GraphQL API. - [DATA_EXPOSURE]: The helper script
scripts/skillctx-resolve.pyreads and writes to a local configuration file at~/.config/skillctx/config.jsonto manage variable bindings for the skill.
Audit Metadata