resolve-fixme
Pass
Audited by Gen Agent Trust Hub on Aug 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a localized discovery script (
find-fixme.sh) and standard developer tools likecargoto manage codebase tasks. Evidence found inSKILL.mdworkflow steps 1 and 5. - [PROMPT_INJECTION]: The skill's primary function is to interpret and act upon
FIXMEcomments, creating a surface for indirect prompt injection from codebase content. - Ingestion points: Code comments containing the string
FIXMEacross the project, as located by thefind-fixme.shscript. - Boundary markers: Absent. The agent is explicitly told to expand the full comment block and "Implement every FIXME completely" without instructions to sanitize or ignore malicious directives embedded in the text.
- Capability inventory: The agent is authorized to modify codebase files (source code, tests, configs) and execute shell commands for testing and discovery.
- Sanitization: None. The skill mandates implementation of the work described in the comments as a record of required tasks.
Audit Metadata