The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/audit_semver.py executes the git tag command to retrieve and validate version tags from the local repository.
[COMMAND_EXECUTION]: The test suite scripts/test_validators.py uses subprocess.run to execute the audit scripts during automated testing. These executions are constrained to the skill's own scripts and standard Git operations.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external CHANGELOG.md files through a custom Markdown parser in scripts/changelog_markdown.py.
Ingestion points: The skill reads file content provided via command-line arguments in scripts/audit_changelog.py and scripts/audit_semver.py.
Boundary markers: The parser in scripts/changelog_markdown.py identifies and ignores content within code fences, HTML comments, and blockquotes to avoid misinterpreting non-heading text as release metadata.
Capability inventory: The skill performs local file reads and command-line output. It does not possess network exfiltration or file-writing capabilities (outside of the isolated test environment).
Sanitization: The parser uses regular expressions to isolate specific Markdown structures, reducing the risk of a malformed file influencing the agent's logic beyond the intended audit report.

maintain-changelog