The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external legal facts and evidence inventories to generate strategy recommendations and pleading outlines.\n
Ingestion points: The scripts/create_strategy_batch.mjs script reads untrusted JSON payloads containing matter details and facts.\n
Boundary markers: The SKILL.md file defines a strict human-in-the-loop boundary, requiring explicit approval for all legal work products before they are exported or used.\n
Capability inventory: The skill has the capability to write to the local file system (scripts/export_strategy_pack.mjs) and access the network via the busabase-sdk to communicate with Busabase bases.\n
Sanitization: Employs deterministic QA checks to identify evidence gaps and unsupported legal positions, mitigating the risk of incorrect agent-generated strategy.\n- [COMMAND_EXECUTION]: The skill instructs the user to execute several local Node.js scripts for management tasks, including scripts/create_strategy_batch.mjs, scripts/execute_decisions.mjs, and scripts/export_strategy_pack.mjs.\n- [EXTERNAL_DOWNLOADS]: The skill depends on external packages such as busabase-sdk, hono, and @hono/node-server downloaded from the official npm registry.\n- [DYNAMIC_EXECUTION]: The skill utilizes dynamic imports to load local configuration files and provider modules at runtime in files like scripts/sync-content.mjs and app/js/providers/index.js.

kelly-legal-matter-strategy