The Agent Skills Directory

[SAFE]: The skill follows a transparent and logical workflow for ticket management. It does not contain any obfuscated code, credential harvesting, or unauthorized persistence mechanisms.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data by fetching context from provided URLs, issue bodies, and comments (Step 1). This creates a potential surface for indirect prompt injection if an attacker embeds instructions in those external sources. However, the risk is mitigated by a mandatory human-review step (Step 4: Quiz the user), where the agent must present the proposed tickets for user approval before any publishing or file-writing occurs.
[COMMAND_EXECUTION]: Codebase exploration and writing to local .scratch directories are standard development workflows. The skill does not attempt to execute arbitrary shell commands or acquire elevated privileges.
[DATA_EXFILTRATION]: Interaction with issue trackers (GitHub, Linear) is the primary intended function of the skill and is performed only after explicit user approval of the content to be published.

to-tickets