The Agent Skills Directory

[SAFE]: The skill's instructions are focused entirely on improving technical documentation style, such as using tables for comparisons, bullet points for facts, and direct phrasing. It includes specific safeguards to ensure the agent preserves technical facts, paths, and commands without invention.
[COMMAND_EXECUTION]: The skill mentions the use of the strictdoc command-line tool (specifically strictdoc export .) to verify documentation syntax. This is consistent with the skill's primary purpose of technical documentation management and does not involve arbitrary or malicious command execution.
[INDIRECT_PROMPT_INJECTION]: The skill inherently processes external data (technical documents like READMEs and specs). The instructions mitigate risks by explicitly directing the agent to maintain the requested scope, preserve machine-readable structures, and avoid changing normative requirements (e.g., keeping "must" as "must"). The threat is considered low and managed by the skill's constraints.

readable-docs