The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute shell commands, including a local script (scripts/check-whitespace.py) and a Python one-liner for YAML validation. The use of placeholders like <target-file-or-dir> and <file> in shell command templates creates a potential command injection surface if the agent platform does not sanitize the user-provided filenames during string interpolation.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection through the following vectors:
Ingestion points: Untrusted data enters the agent context via the Read tool when processing target markdown files in Phase 2.
Boundary markers: The instructions lack delimiters or explicit directives to ignore instructions embedded within the target files.
Capability inventory: The skill has access to Edit, Write, and Bash tools, which could be exploited if the agent follows malicious instructions from the file content.
Sanitization: No sanitization or filtering is performed on the ingested content; the instruction to "KEEP (never cut): Every instruction" specifically mandates the preservation of all instruction-like content, including potential adversarial injections.

condense