The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the GitHub CLI (gh) to perform actions such as listing, viewing, creating, and editing issues and labels. These operations are strictly controlled through a multi-phase orchestration protocol that requires explicit user approval before any write actions are executed (Evidence: references/orchestration-protocol.md Phase 7).
[DATA_EXFILTRATION]: The skill reads local repository source code and existing GitHub issue content to perform its investigation and grouping tasks. While it engages in network operations to interact with GitHub, these actions are essential to its primary function and target a well-known, trusted service.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted data from external sources, specifically user-provided bug reports and existing GitHub issue descriptions. This data is used by sub-agents to generate new issue content. The skill lacks explicit boundary markers or sanitization for this ingested data.
Ingestion points: User reports in SKILL.md (Phase 0/1) and existing GitHub issue bodies in references/orchestration-protocol.md (Phase 2/6).
Boundary markers: None identified; the skill does not use specific delimiters or instructions to ignore embedded commands within user input or external issue text.
Capability inventory: The skill has the ability to write to the repository via gh issue create and gh issue edit.
Sanitization: No escaping or filtering is applied to external content before it is processed by the AI agents.

issue-report