context-firewall

Run the context-firewall workflow: split inputs, invoke FileWorkers, aggregate into SubResult.v1.

Generate a TaskSpec.v1 JSON template for the context-firewall workflow.

Verify SubResult.v1 by sampling evidence locators and re-reading only referenced snippets.

Use this agent when multiple FileWorker results need to be merged into a single SubResult.v1, with de-duplication, clustering, and conflict detection.

Use this agent when a high-risk TaskSpec/SubResult needs an independent second pass to reduce hallucinations, validate coverage, or resolve cross-file causal claims.

Use this agent when a task requires reading or scanning large inputs (files/logs/PDFs/images/tool outputs) and returning a compressed, auditable SubResult fragment with evidence locators.

This skill should be used when the user asks to "analyze a large log", "summarize a big file", "read the whole repository", "process long tool output", "avoid context window overflow", "use sub-agents for preprocessing", "add evidence with line numbers", "verify claims with locators", or mentions "TaskSpec", "SubResult", "Evidence Contract", "Context Firewall", "Map-Reduce", "FileWorker", "Aggregator", "Critic".

4 hooks across 3 events