The Agent Skills Directory

[DATA_EXFILTRATION]: The skill references a file path ../../../artifacts/changelog-import-2026-04-05.txt in SKILL.md, which involves accessing content outside the primary skill directory.
[DATA_EXFILTRATION]: The skill includes instructions to transmit generated content to an external Discord webhook using curl.
[COMMAND_EXECUTION]: The workflow relies on executing shell commands through the gh CLI to retrieve repository metadata and curl for network delivery.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes untrusted input from GitHub pull request descriptions and titles.
Ingestion points: PR titles and descriptions fetched via the gh CLI in references/full-guide.md.
Boundary markers: No delimiters are specified to isolate untrusted data.
Capability inventory: Shell command execution via gh and network transmission via curl.
Sanitization: No sanitization or filtering of external PR content is specified before it is included in the output.

changelog