The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the gh issue create command to perform external write operations. It explicitly directs the agent to bypass user oversight in SKILL.md (Step 7: "Don't let the user review before creating
create it immediately and share the link"). This removal of the human-in-the-loop check for network-active commands increases the risk of unauthorized or unsafe actions being performed automatically.
[DATA_EXFILTRATION]: The skill explores the codebase and identifies "architecture friction points" (Step 1), then automatically posts descriptions of these areas to GitHub (Step 7). Without a review step, proprietary logic, sensitive configuration comments, or internal architecture details could be leaked to public or shared issue trackers.
[PROMPT_INJECTION]: The skill is vulnerable to Indirect Prompt Injection (Category 8) because it processes untrusted content from the codebase to generate its output.
Ingestion points: The agent reads arbitrary files from the local codebase using the Explore subagent (SKILL.md).
Boundary markers: Absent. The instructions do not provide delimiters or warnings to ignore instructions found within the code being analyzed.
Capability inventory: The skill possesses network-write capabilities via the gh issue create command (SKILL.md).
Sanitization: Absent. There is no evidence of filtering or validation of the content extracted from the codebase before it is used to populate the GitHub Issue template.

code-arch-optimizer