The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted external data from X (authored posts and bookmarks) to influence its drafting behavior.
Ingestion points: The agent is instructed to open https://x.com/search and https://x.com/i/bookmarks to collect text content from external users.
Boundary markers: The skill contains a directive to "Separate reusable patterns from one-off artifacts such as pasted prompts," which serves as a basic instruction-level defense but does not provide a robust structural boundary.
Capability inventory: The skill possesses the capability to write files to the local workspace (data/x-growth/bookmark-quote-posts/) and perform Git operations (git commit).
Sanitization: There is no evidence of programmatic sanitization or escaping of the ingested X content before it is processed by the model for drafting.
[COMMAND_EXECUTION]: The skill uses shell-based version control commands to manage the workspace state.
Evidence: SKILL.md explicitly calls for the execution of git status --short, git diff --check, and git commit to track changes and save the generated drafts.

x-bookmark-quote-posts