The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes several command-line utilities including git, gh (GitHub CLI), glab (GitLab CLI), and curl to automate the repository detection and issue submission process. It also executes runtime version checks for various programming languages to gather environment information for the report.
[DATA_EXFILTRATION]: By design, the skill transmits project-related information such as bug descriptions, code snippets, and media attachments to external platforms (GitHub or GitLab). This risk is managed through a multi-layered defense strategy: a dedicated privacy sanitization phase that redacts sensitive paths and credentials, and a mandatory 'HARD STOP' review where the user must approve the final sanitized content before submission.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the active conversation history and the local codebase to generate issue drafts, creating a potential surface for indirect prompt injection.
Ingestion points: File contents accessed via Read, Glob, and Grep tools, session context reconstruction, and user-provided arguments.
Boundary markers: Uses structured templates to separate different parts of the issue report, although data is interpolated into shell commands for submission.
Capability inventory: Shell execution capabilities via gh, glab, and curl for platform interaction.
Sanitization: Implements a mandatory threat scan (Step 9.5) and privacy sanitization protocol (Step 10) to filter and redact malicious or sensitive content before it reaches the platform.

qa-issue-report