badger-triage
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the GitHub CLI (
gh api graphql,gh issue edit) to interact with GitHub Projects and Issues. These commands are used to query project structure and apply updates (milestones, sizes, priorities) to the project board. The instructions emphasize human-in-the-loop confirmation before any write operations are executed. - [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources, specifically GitHub issue titles and labels from the
AutumnsGrove/Latticerepository. This constitutes a potential injection surface if issue content contains malicious instructions. However, the risk is significantly mitigated by the mandatory 'Discuss' phase, where the agent presents findings to the user and awaits confirmation before taking action. - [CREDENTIALS_UNSAFE]: The
references/github-graphql.mdfile contains several hardcoded identifiers such as Project IDs (PVT_kwHOAiMO684BNUxo) and Field IDs (PVTSSF_lAHOAiMO684BNUxozg8WnH4). These are standard GitHub GraphQL node identifiers used to target specific project fields and do not represent sensitive secrets or API tokens.
Audit Metadata