The Agent Skills Directory

[SAFE]: The skill provides authoritative guidance based on official documentation. Code examples for PHP and JavaScript follow standard TYPO3 extension development practices and security principles.
[REMOTE_CODE_EXECUTION]: Automated scanners flagged several commands where curl output is piped to python3. Manual inspection confirms these are safe debugging commands that use 'python3 -m json.tool' to format JSON responses from a local Solr instance. They do not execute remote content as code.
Evidence in SKILL.md: ddev exec curl -s http://solr:8983/solr/core_en/admin/ping | python3 -m json.tool and similar diagnostic commands.
[COMMAND_EXECUTION]: Includes instructions for using CLI tools (curl, composer, ddev) to manage local infrastructure. These operations are restricted to the local development environment or internal network services (e.g., solr:8983, tika:9998).
[EXTERNAL_DOWNLOADS]: References official software repositories and documentation from trusted organizations (Apache, TYPO3, DDEV). No downloads from untrusted or high-risk third-party sources were identified.
[PROMPT_INJECTION]: Evaluated for indirect prompt injection vulnerability with a low risk profile.
Ingestion points: The skill ingests user search queries and indexes content from TYPO3 records and files (PDF, DOCX) via EXT:solr and EXT:tika.
Boundary markers: Relies on standard TYPO3 indexing logic; specific LLM boundary markers are not explicitly defined in the technical setup.
Capability inventory: The skill facilitates search indexing and retrieval but does not expose arbitrary command execution or file system access to the search input.
Sanitization: Employs the SOLR_CONTENT object to sanitize indexed data by stripping HTML and RTE formatting.

typo3-solr