The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/validate_skill_project.py script uses subprocess.run to invoke bash -n for non-executing syntax validation of shell scripts. The scripts/verify_npx_install.sh script also uses npx to verify that skills can be correctly installed via the platform's package manager.
[DYNAMIC_EXECUTION]: The scripts/validate_skill_project.py script utilizes the Python compile() function to perform syntax checks on newly generated Python files. The code is compiled to verify correctness but is not executed by the validator.
[EXTERNAL_DOWNLOADS]: The skill instructions and scripts suggest using npx to install the skills utility and add the author's own package (dontbesilent2025/dbskill). This is a legitimate dependency for the skill's stated purpose of managing agent skills.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it ingests user-provided descriptions and goals to generate new skill files.
Ingestion points: User input provided to scripts/init_skill_project.py and requirement analysis workflows in references/problem-and-goal.md.
Boundary markers: Not explicitly defined in the generation templates.
Capability inventory: File system writes, subprocess execution for validation, and network operations via npx and git.
Sanitization: The scripts perform basic string whitespace normalization but no complex sanitization of the generated prompt instructions.

dbs-skill-maker