The Agent Skills Directory

[DYNAMIC_CONTEXT_INJECTION]: The skill uses the !command syntax in SKILL.md to execute shell commands at load time. Specifically, it runs git log and git tag to populate the prompt with recent history and executes a local helper script (yaml-helper.sh) to resolve configuration keys. These commands are used for legitimate context gathering and do not involve network operations or sensitive file access.
[INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted content from git commit messages, sprint reports, and design documents, which could contain malicious instructions designed to influence the agent's behavior during changelog generation.
Ingestion points: Git commit subjects/logs and files within production/sprints/ and design/gdd/ folders.
Boundary markers: The skill does not employ explicit delimiters or instructions to ignore embedded commands within the processed data.
Capability inventory: The skill allows access to Write, Edit, and restricted Bash tools in SKILL.md, enabling the agent to modify project files (e.g., docs/CHANGELOG.md) based on its analysis.
Sanitization: There is no evidence of sanitization, escaping, or filtering of the content read from external files before it is processed by the model.
[COMMAND_EXECUTION]: Shell command execution is limited to git utilities for repository introspection and a specific local configuration script as defined in the allowed-tools frontmatter. These tools are used for their intended purpose of gathering change data.

write-changelog