ollama_change_brief

Change brief. FLAGSHIP compound job. Produces a CHANGE IMPACT BRIEF: {change_summary, affected_surfaces, why_it_matters, likely_breakpoints, validation_checks, release_note_draft, evidence, weak, coverage_notes, corpus_used}. Accepts diff_text (split per file on diff --git markers) and/or source_paths (changed files), with an optional corpus for architecture context. Corpus-sourced evidence items carry retrieval score. Optional corpus_min_evidence_score drops corpus chunks below the relevance floor before the model sees them. Not a git chat bot — structured and reviewable. likely_breakpoints are INVESTIGATIVE reasoning about what could break; validation_checks are what to verify after the change. Never remedial (no ‘apply this fix’). release_note_draft is a draft the operator reviews. If you pre-extracted claims via ollama_extract with a frame argument, drop off-topic items before assembling source_paths/diff_text — there’s no in-brief topicality gate for diff/path inputs.

Parameters

Parameter	Type	Required	Default	Description
diff_text	string	no	—	Unified-diff text (e.g. git diff output). Split per file on diff --git markers into numbered evidence items.
source_paths	string[]	no	—	Changed files to read server-side. Use alongside diff_text when the full file context matters, or alone when no diff is available. Optional — diff-driven calls work without it; runtime requires at least one of diff_text or source_paths.
corpus	string	no	—	Optional: named corpus (e.g. ‘handbook’, ‘doctrine’) for architecture and release-process context.
corpus_query	string	no	—	Query used to pull chunks from the corpus. Defaults to a digest of the diff/path heads.
per_file_max_chars	integer	no	—	Chars per source file (default 20k).
max_breakpoints	integer	no	—	Cap on likely_breakpoints (default 6).
max_validation_checks	integer	no	—	Cap on validation_checks (default 8).
corpus_min_evidence_score	number	no	—	Minimum retrieval score (0–1) for a corpus chunk to enter evidence. Hits below the floor are dropped before the model sees them, with a counted note in coverage_notes. Absent → no relevance filter.

Full input schema

The JSON Schema below is generated from the same zod schema the server validates against at the wire — it cannot drift from runtime behavior.

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "type": "object",
  "properties": {
    "diff_text": {
      "description": "Unified-diff text (e.g. `git diff` output). Split per file on `diff --git` markers into numbered evidence items.",
      "type": "string",
      "minLength": 1
    },
    "source_paths": {
      "description": "Changed files to read server-side. Use alongside diff_text when the full file context matters, or alone when no diff is available. Optional — diff-driven calls work without it; runtime requires at least one of diff_text or source_paths.",
      "type": "array",
      "items": {
        "type": "string",
        "minLength": 1
      }
    },
    "corpus": {
      "description": "Optional: named corpus (e.g. 'handbook', 'doctrine') for architecture and release-process context.",
      "type": "string",
      "pattern": "^[a-zA-Z0-9_-]+$"
    },
    "corpus_query": {
      "description": "Query used to pull chunks from the corpus. Defaults to a digest of the diff/path heads.",
      "type": "string",
      "minLength": 1
    },
    "per_file_max_chars": {
      "description": "Chars per source file (default 20k).",
      "type": "integer",
      "minimum": 1000,
      "maximum": 200000
    },
    "max_breakpoints": {
      "description": "Cap on likely_breakpoints (default 6).",
      "type": "integer",
      "minimum": 1,
      "maximum": 12
    },
    "max_validation_checks": {
      "description": "Cap on validation_checks (default 8).",
      "type": "integer",
      "minimum": 1,
      "maximum": 15
    },
    "corpus_min_evidence_score": {
      "description": "Minimum retrieval score (0–1) for a corpus chunk to enter evidence. Hits below the floor are dropped before the model sees them, with a counted note in coverage_notes. Absent → no relevance filter.",
      "type": "number",
      "minimum": 0,
      "maximum": 1
    }
  }
}

Envelope

Every tool returns the standard envelope — tier_used, model, tokens_in/tokens_out, elapsed_ms, backend provenance, warnings. See Envelope & tiers.

Source

Schema + handler: src/tools/changeBrief.ts · registration: src/index.ts

---

Back to the Tool Reference.