The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes a bash script scripts/check-prose.sh that utilizes standard Unix utilities like awk, grep, and sed to search for specific linguistic patterns in text files. It also provides a Python one-liner in SKILL.md for analyzing sentence length.
[DYNAMIC_EXECUTION]: The skill provides a Python snippet intended to be run by the user or agent to calculate text statistics. This is a local script provided as part of the skill documentation.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process user-supplied prose drafts, which represents a potential attack surface for indirect prompt injection.
Ingestion points: Untrusted prose drafts are processed via the FILE argument in the Python one-liner (in SKILL.md) and the scripts/check-prose.sh script.
Boundary markers: Absent. The instructions do not define delimiters or treat the input file content as potentially untrusted data.
Capability inventory: Local command execution using bash, awk, and grep (in scripts/check-prose.sh), and Python execution (in SKILL.md).
Sanitization: Absent. The analysis tools operate directly on the input text using regular expressions without prior sanitization.

public-writing