The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill encourages the collection and storage of sensitive credentials in plaintext documentation files.
Evidence: In 'Phase 1: Gather App Information', the skill requests 'Developer account details (Apple ID, Team ID)' and 'Demo account credentials'.
Evidence: In 'Phase 2.2: Marketing Copy', instructions include generating 'Review Notes' that contain 'Demo account credentials'.
Risk: Writing these secrets directly into Markdown files (e.g., Master Submission Guide) makes them accessible to anyone with access to the repository, increasing the risk of account compromise.
[INDIRECT_PROMPT_INJECTION]: The skill serves as a processing engine for untrusted user-provided data without adequate safeguards.
Ingestion points: The skill collects extensive user input in 'Phase 1', including app names, feature descriptions, and unique selling points.
Boundary markers: Absent. The instructions do not define delimiters or provide 'ignore embedded instructions' warnings for the interpolated data.
Capability inventory: The skill utilizes file-writing capabilities to generate at least 8 documents in the /docs/ directory and creates system checkpoints.
Sanitization: Absent. There is no requirement for the agent to sanitize or escape HTML/Markdown characters in user-provided strings before writing them to the templates, which could lead to XSS or downstream injection if the docs (like the Support Page) are rendered in a browser.

app-store-submission-packager