forgeplan_blindspots

Show blind spots in the workspace — active decisions (PRD, RFC, ADR, Epic) that have no linked evidence, plus orphan artifacts with zero incoming or outgoing links. Blind spots are the #1 technical debt signal in Forgeplan: an active PRD with no evidence is a false promise.

Category: Dashboards & Graph

When an agent calls it

After session start — if forgeplan_health reports verdict: debt, call this for the detailed list.
Before merging a PR — make sure new work doesn’t leave new blind spots.
During refactoring — find orphan Notes / Problems that should be deprecated or re-linked.
Quality sweeps — produce a fix list sortable by owner / kind.

Input parameters

No input parameters. Call this tool with an empty object {}.

Returns

{
  "blind_spots": [
    {
      "id": "PRD-042",
      "kind": "prd",
      "status": "active",
      "reason": "no linked evidence",
      "r_eff": 0.0
    },
    {
      "id": "RFC-006",
      "kind": "rfc",
      "status": "active",
      "reason": "no linked evidence"
    }
  ],
  "orphans": [
    {
      "id": "NOTE-017",
      "kind": "note",
      "reason": "no incoming or outgoing links"
    }
  ],
  "summary": {
    "blind_spots_count": 2,
    "orphans_count": 1
  }
}

Example invocation

{}

Typical sequence

forgeplan_blindspots — list offenders.
For each blind spot: forgeplan_new evidence → forgeplan_link → forgeplan_score.
For each orphan: either forgeplan_link to a parent, or forgeplan_deprecate.
Re-run forgeplan_health → expect verdict: healthy.

CLI equivalent

forgeplan blindspots