The Agent Skills Directory

[DYNAMIC_CONTEXT_INJECTION]: The skill uses the ! syntax in SKILL.md to execute local shell scripts (yaml-helper.sh and story-status.sh) at the moment the skill is loaded. These scripts gather project configuration and current story statuses to provide immediate context to the agent. These scripts are located within the project's own .claude directory structure.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes data from multiple external project files to generate its report. This creates a surface where malicious content in those files could influence the agent's behavior.
Ingestion points: production/sprints/ markdown files, production/sprint-status.yaml, and story files matching the glob production/epics/**/story-*.md.
Boundary markers: Absent; the skill does not use specific delimiters or instructions to ignore potential commands embedded in the processed files.
Capability inventory: The skill has access to Bash (script execution), Read, Glob, and Grep tools.
Sanitization: None; the content of the sprint and story files is read and processed without filtering or validation of the text content.
[COMMAND_EXECUTION]: The skill is configured to allow the agent to execute specific bash scripts via the allowed-tools frontmatter. While these are restricted to project-specific paths (.claude/scripts/), one configuration (Bash(bash .claude/scripts/story-status.sh*)) uses a wildcard that could potentially permit the execution of the script with unexpected or malicious arguments if the model is influenced by poisoned file content.

sprint-status