The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest untrusted data from Obsidian .base files (YAML) and markdown frontmatter to generate SQL queries.
Ingestion points: The skill processes .base files containing filters, formulas, and views (SKILL.md).
Boundary markers: There are no instructions for the agent to use boundary markers or delimiters when interpolating data from these files into SQL strings.
Capability inventory: The generated SQL is intended for execution via the sense tool, which has access to the local filesystem and metadata database.
Sanitization: The skill provides direct translation mappings (e.g., contains(x) to instr(f.string, x) > 0) without explicitly requiring the agent to escape or parameterize user-controlled values, creating a potential SQL injection surface.
[DYNAMIC_EXECUTION]: The skill facilitates the generation of executable SQL code from user-supplied templates.
Evidence: SKILL.md and EXAMPLES.md provide logic for constructing complex SQL queries (including Common Table Expressions and subqueries) by concatenating strings derived from user-provided YAML configurations. The generated code is subsequently executed by the sense CLI.

sense-bases