The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes several scripts designed to automate developer workflows by executing local system commands.
scripts/setup-typo3-coredev.sh uses git, ddev, and docker to set up the local development environment.
scripts/validate-commit-message.py uses subprocess.run to execute git log and retrieve the most recent commit message for validation.
[INDIRECT_PROMPT_INJECTION]: The skill facilitates interaction with external project management platforms (TYPO3 Forge and GitLab), creating an attack surface where untrusted data could influence the agent.
Ingestion points: Data is fetched from the TYPO3 Forge REST API (scripts/query-forge-metadata.sh) and the TYPO3 GitLab REST API (scripts/t3o-gitlab.py).
Boundary markers: The scripts do not explicitly use delimiters or instruction-guarding prompts when displaying external content to the agent.
Capability inventory: The skill provides capabilities to write and push code (git commit, git push) and interact with APIs to create/update issues and merge requests.
Sanitization: scripts/t3o-gitlab.py implements proactive security by validating URL schemes to ensure only http and https are used, preventing potential local file disclosure via file:// URIs.
[PRIVILEGE_ESCALATION]: Documentation within the skill suggests practices that result in overly permissive file system settings.
Evidence: references/ddev-setup-workflow.md recommends using chmod -R 777 on project directories (var/, typo3temp/, typo3conf/) as a troubleshooting step for write permission errors, which is an insecure configuration.

typo3-core-contributions