forgeplan_calibrate

Suggest an appropriate depth level (Tactical / Standard / Deep / Critical) for one or all artifacts, based on heuristic content analysis. Depth drives validator strictness (MUST rules per depth), ADI requirement, and review ceremony — so miscalibrated depth means either over-engineering a one-liner or shipping a critical change without adversarial review.

Category: Quality

When an agent calls it

After creating a stub — let calibration suggest the right depth before filling in sections.
During review — sanity-check that the author chose a reasonable depth.
Audit sweep — find artifacts marked Tactical that actually warrant Deep based on body content.
Onboarding agents — teach them what factors push depth up (security, breaking changes, cross-team scope).

Heuristics considered: security sections present, breaking-change markers, outgoing link count, body length and section complexity.

Input parameters

Name	Type	Required	Description
`id`	`string`	no	Artifact ID to calibrate. Omit to calibrate all artifacts in the workspace.

Schema source: crates/forgeplan-mcp/src/server.rs::CalibrateParams

Returns

{
  "suggestions": [
    {
      "id": "PRD-042",
      "current_depth": "tactical",
      "suggested_depth": "deep",
      "reasons": [
        "Security section present",
        "Breaking change detected (## Migration required)",
        "8 outgoing links"
      ]
    }
  ],
  "total_suggestions": 1
}

Example invocation

{ "id": "PRD-042" }

Typical sequence

forgeplan_calibrate — scan the workspace.
For each mismatch: re-route the task with forgeplan_route or manually set depth in frontmatter.
forgeplan_validate — rerun validation under the new depth (stricter MUST rules).
forgeplan_reason — ADI is mandatory at Deep / Critical.

CLI equivalent

forgeplan calibrate
forgeplan calibrate PRD-042