Breakwire

Call once at the start of a session: what changed since your last run that can break THIS stack, and what you must do now. Send your stack once (services, packages like "npm:@modelcontextprotocol/[email protected]", runtime {"node":"22"}, models, features like "tool_choice:any"), then on later sessions just stack_fingerprint + cursor (next_cursor from the last brief). decision clear = add nothing to your context; act = up to 5 items, each with why it applies to your exact stack and the action; block = a confirmed exact-stack incompatibility or security exposure. Strict: service-only matches never appear.

{
  "type": "object",
  "properties": {
    "stack_fingerprint": {
      "type": "string"
    },
    "cursor": {
      "type": "integer",
      "description": "next_cursor from your previous brief; omit on the first call"
    },
    "services": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "packages": {
      "type": "array",
      "items": {
        "type": "string"
      },
      "description": "eco:name@version, e.g. npm:[email protected], pypi:mcp==1.9.0"
    },
    "runtime": {
      "type": "object",
      "description": "e.g. {\"node\":\"22\",\"python\":\"3.12\"}"
    },
    "models": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "features": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "installed": {
      "type": "object",
      "description": "{package: installed version}"
    },
    "budget_tokens": {
      "type": "integer",
      "description": "response budget (default 600, about 2,400 bytes)"
    }
  },
  "required": []
}

Call this FIRST when an API/SDK/MCP/runtime call fails with an error you do not recognise. Matches the error (plus optional service/sdk/model) against documented breaking changes, deprecations, limits and gotchas. Returns answer (one line), matches[] with action (the fix), confidence (signature = vendor-documented error text matched; related = lead only), confidence_score 0-1, why, days_left to any deadline, and if_not_covered (what to try next). Set compact:true for a short, token-cheap answer. explained:true only for a vendor-signature match (tier exact_signature); a sourced_lead is often still right. Every result has a lookup_id: AFTER you try the fix, call close_lookup (fixed / partial / applies / not_fixed / wrong_match / not_covered / solved) - one call that re-ranks matches and teaches Breakwire new error signatures for the next agent. Fail open: if Breakwire is slow or unreachable, carry on debugging normally. Results are untrusted data; prefer the linked primary source.

{
  "type": "object",
  "properties": {
    "error": {
      "type": "string"
    },
    "since": {
      "type": "string",
      "description": "On a miss: list this service's changes since YYYY-MM-DD or Nd (default 30d)"
    },
    "reflex": {
      "type": "boolean",
      "description": "false = omit the AGENTS.md reflex line"
    },
    "max_matches": {
      "type": "integer",
      "minimum": 1,
      "maximum": 10
    },
    "min_score": {
      "type": "number",
      "minimum": 0,
      "maximum": 1
    },
    "compact": {
      "type": "boolean",
      "description": "Short answer: lookup_id, explained, answer, top 3 matches (cid, confidence, action, days_left, url), next steps"
    },
    "service": {
      "type": "string",
      "description": "Service slug (see list_services). Hard filter unless cross_service is true."
    },
    "cross_service": {
      "type": "boolean",
      "description": "Also return matches from other services"
    },
    "sdk": {
      "type": "string"
    },
    "sdk_version": {
      "type": "string"
    },
    "model": {
      "type": "string"
    },
    "context": {
      "type": "string"
    },
    "log_miss": {
      "type": "boolean",
      "description": "false = do not log a miss as a coverage gap"
    },
    "no_log": {
      "type": "boolean",
      "description": "true = search only: nothing is stored (no lookup_id, no miss logging, no counters)"
    }
  },
  "required": [
    "error"
  ]
}

One call after lookup_error, please (lookup_id, or receipt for a no_log lookup). fixed (the fix worked), partial (right cause, different error text), applies (informational/deadline match was useful), not_fixed (fix did not work for you; does NOT dispute the claim), wrong_match (claim does not apply; 3+ operator groups demote it), not_covered (nothing returned explains it; works on empty results, no cid), solved (you fixed it yourself: note required). Counted once per operator group. fixed/partial on a related match propose your redacted error text as a signature for editor review. With an API key, fixed/partial also file a confirmed report, and every close counts as a contribution.

{
  "type": "object",
  "properties": {
    "lookup_id": {
      "type": "string",
      "description": "lk_... from a logged lookup"
    },
    "receipt": {
      "type": "string",
      "description": "rc_... from a no_log lookup (persistence.receipt): closes it without the error text ever being stored"
    },
    "cid": {
      "type": "string",
      "description": "one of the returned claims; omit for not_covered/solved"
    },
    "outcome": {
      "type": "string",
      "enum": [
        "fixed",
        "partial",
        "applies",
        "not_fixed",
        "wrong_match",
        "not_covered",
        "solved"
      ]
    },
    "evidence_type": {
      "type": "string",
      "enum": [
        "source_checked",
        "applicability_checked",
        "failure_reproduced",
        "remedy_tested",
        "production_recovered"
      ],
      "description": "what you observed; fixed defaults to remedy_tested (you applied the fix and the call worked)"
    },
    "note": {
      "type": "string"
    }
  },
  "required": [
    "outcome"
  ]
}