The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill contains a hardcoded IndexNow verification key 'f1e4781767e4472e9061ad0f853449d3' in Step 9. Hardcoding authentication keys or verification secrets in instruction files is a risk to security best practices.- [COMMAND_EXECUTION]: The skill utilizes several shell commands to interact with the local environment, including git for repository management, pnpm for building project components, and python3 for generating JSON payloads via heredoc strings. These operations provide significant control over the local execution environment.- [EXTERNAL_DOWNLOADS]: The skill uses curl to make external network requests to qaskills.sh for status verification and to search engine endpoints www.bing.com and api.indexnow.org to submit content updates.- [PROMPT_INJECTION]: The skill demonstrates a surface for indirect prompt injection by processing untrusted data from WebSearch results and user-provided keywords in Step 1. Ingestion points: Keyword lists and WebSearch results used as the basis for article generation. Boundary markers: No specific delimiters or safety instructions are provided to the agent to distinguish between data and instructions during content generation. Capability inventory: The skill has access to the local filesystem, build tools, Git, and network operations via curl and python3 scripts. Sanitization: Instructions focus on technical escaping for template literals (e.g., backticks and interpolation symbols) but do not address semantic safety of the ingested data.

publish-seo-batch