The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze potentially untrusted content from URLs, local files, or direct user input to generate social media hooks. This represents a vulnerability surface where malicious instructions embedded in the source text could attempt to influence the agent's output.
Ingestion points: Text is loaded from remote URLs, local file paths, or user-provided content following a --- delimiter.
Boundary markers: The workflow uses a --- separator line to demarcate the source text from the generation instructions.
Capability inventory: The skill performs file read operations and executes local scripts (last-article.sh and check_hooks.py).
Sanitization: There is no programmatic sanitization of the input text; the skill relies on instructional guardrails provided to the agent.
[COMMAND_EXECUTION]: The skill executes local shell and Python scripts as part of its internal workflow. It calls ${CLAUDE_SKILL_DIR}/../write-review/scripts/last-article.sh to identify the most recent relevant content and uses scripts/check_hooks.py to perform automated validation of the generated social media hooks.

write-hook