The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from external technical sources and repository files to automate actions, creating a surface for indirect instructions to influence agent output.
Ingestion points: The skill is instructed to fetch content from https://jser.info/ and read various repository files, including package.json, project reports in .frontend-review/, and markdown files.
Boundary markers: There are no explicit instructions or delimiters provided to ensure the agent treats the content from the external technical blog or internal report files as potentially untrusted data.
Capability inventory: The skill has the capability to write files to the repository, create GitHub issues using the gh CLI, and dispatch multiple specialized agent skills.
Sanitization: The instructions do not define any validation or sanitization steps for the external data before it is used to generate reports or file issues.
[EXTERNAL_DOWNLOADS]: The skill performs network operations to retrieve technical updates from an external domain.
Evidence: The skill uses WebFetch to access https://jser.info/ to monitor package releases and security advisories for the project's dependencies.
[COMMAND_EXECUTION]: The skill automates the creation of repository issues based on its findings.
Evidence: The instructions explicitly direct the agent to use gh issue create to track regressions and findings from the weekly review process.

frontend-review-weekly