The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill operates by ingesting and analyzing external database queries and execution plans, creating a surface for potential indirect prompt injection attacks.
Ingestion points: The skill instructions and workflows in SKILL.md and references/query-optimization.md involve processing untrusted input in the form of query strings and EXPLAIN plan outputs.
Boundary markers: The skill lacks explicit instructions or delimiters that would prompt the agent to treat these external inputs as potentially untrusted data.
Capability inventory: The skill documentation (e.g., postgresql-tuning.md, mysql-tuning.md) provides templates for high-impact database management operations, including ALTER SYSTEM, CREATE INDEX CONCURRENTLY, ALTER TABLE, and VACUUM FULL, which could be exploited if an injection occurs.
Sanitization: There are no provisions for sanitizing or filtering external database content before it is processed by the agent.
[NO_CODE]: The skill is composed entirely of Markdown documentation and SQL templates; it does not include scripts, executable binaries, or external package dependencies, and references documentation on a well-known service.

database-optimizer