The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by ingesting untrusted data from external sources.
Ingestion points: External content is fetched in Phase 2 of SKILL.md using the gh CLI and WebFetch to read GitHub PRs, issues, and discussions.
Boundary markers: The instructions lack explicit boundary markers or directives for the agent to ignore instructions that might be embedded within the fetched external content.
Capability inventory: The agent has the capability to write to persistent memory using mcp__serena__write_memory (Phase 6) and conduct further external research via WebSearch or Perplexity tools.
Sanitization: There is no evidence of sanitization or validation of the external content before it is processed or stored in memory.
[COMMAND_EXECUTION]: The skill uses the GitHub CLI (gh) to retrieve structured data about issues and pull requests. These commands are used legitimately to fulfill the skill's purpose of technical analysis.
[EXTERNAL_DOWNLOADS]: Technical content and documentation are fetched from well-known and trusted platforms including GitHub (github.com), the V8 engine blog (v8.dev), and MDN (developer.mozilla.org).

core-topic