The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses a Python script to perform word count validation on generated output files.
Evidence: python3 skills/shared/scripts/wordcount.py count -f outputs/主题名/copy.txt in SKILL.md (Step 9).
Context: This command is used to ensure marketing copy meets platform-specific character limits. The file path includes a variable placeholder (主题名) which may be derived from user input or task metadata.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input and incorporates it into file paths used in shell commands.
Ingestion points: User prompt input (product details, audience, subject name) described in SKILL.md (Step 1).
Boundary markers: Absent. The skill does not explicitly instruct the agent to sanitize input before using it in the word count command.
Capability inventory: Subprocess execution via python3 (Step 9 in SKILL.md) and file write operations to the outputs/ directory.
Sanitization: Absent. There are no instructions for escaping or validating the 'Subject Name' used in the output file path, which could lead to path traversal or command injection if the agent does not handle the string safely.

copywriting