wayfinder
Pass
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from an external issue tracker (e.g., GitHub, GitLab, or local Markdown files) which serves as a vulnerability surface for malicious instructions.
- Ingestion points: As defined in
SKILL.md, the agent loads the 'Map' (an issue with thewayfinder:maptag) and individual 'Tickets' (sub-issues), including their titles and body content, from the issue tracker. - Boundary markers: The instructions do not define specific delimiters or "ignore instructions" warnings to help the agent distinguish its own system instructions from potentially malicious commands embedded within the retrieved issue content.
- Capability inventory: The skill has broad capabilities, including reading and writing issues, creating git branches (
research/<name>), and spawning sub-agents for research tasks as detailed in the 'Advancing the Map' section ofSKILL.md. - Sanitization: The skill does not mention or implement sanitization or validation of the content retrieved from the issue tracker before it is processed by the agent's core reasoning engine.
Audit Metadata