fusion-issue-solving
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill is designed to execute validation commands and project checks within the local repository environment to verify implementation work.
- Evidence: Instruction 8 in
SKILL.mdexplicitly directs the agent to "Run targeted checks first, then required project checks." - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external GitHub issues, which represents a potential surface for indirect instructions embedded in issue bodies or discussions.
- Ingestion points:
SKILL.mdInstruction 2 requires the agent to read the issue body, labels, and linked discussions. - Capability inventory: The agent has the ability to write local files (
.tmp/directory), execute shell commands for validation, and perform GitHub API mutations. - Boundary markers: While explicit technical delimiters are not defined, the workflow requires the agent to "Confirm issue context and success criteria" with the user before proceeding (Instruction 2).
- Sanitization: No specific sanitization or filtering of the issue content is mentioned, relying instead on the confirmation step.
Audit Metadata