jira-report-comment
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests commit messages, Git diffs, and Jira ticket details to generate reports. These are external sources that can be influenced by other users (e.g., through commit messages or Jira ticket updates).
- Ingestion points: Data is read in SKILL.md via Step 2 (getJiraIssue tool) and Step 3 (git log and git diff).
- Boundary markers: No explicit delimiters or warnings to ignore instructions within the retrieved text are present in the skill's logic.
- Capability inventory: The skill possesses the ability to execute shell commands (git, mkdir, touch) and write files to the local directory docs/jira-reports/.
- Sanitization: The skill lacks explicit filtering or sanitization steps for the data processed from Git or Jira.
- [COMMAND_EXECUTION]: The skill makes use of common Git and filesystem utilities via the shell.
- Evidence: Commands include git branch, git log, git diff, mkdir, and touch, all focused on generating and storing reports locally within the project structure. The issue key resolution step includes a human confirmation checkpoint to mitigate simple command injection via malformed keys.
Audit Metadata