@javierfrancoleal/accountability-auditor 1.0.0 → 1.1.1

package/assets/accountability-auditor.md

@@ -4,6 +4,17 @@ Final stage of the J.Franco agentic workflow:
 
 `Input → Detection (AI + human) → AI processes, decides, and proposes → Human reflects, verifies, and decides → Output`
 
+## Crossed check (folded from the cortex organ)
+
+The hemisphere that proposes is never the hemisphere that verifies:
+
+- Every proposal is checked once from the opposite angle (patterns and
+  risks, not the same reasoning that produced it).
+- Before presenting a proposal, predict the expected outcome in one
+  sentence; if reality diverged, report the surprise explicitly.
+- Attach the crossed verdict in one line ("Crossed check: …") to the
+  Accountability Report.
+
 ## Purpose
 
 Prepare the human's verification and decision before any output is delivered
@@ -97,6 +108,12 @@ and the human's decisions. This file is the owner's living record of what
 the AI did in the company; never edit or remove earlier entries. The
 `autonomy-grader` skill reads this log to propose tier moves.
 
+**Usage reporting:** when a supervisor is reachable
+(http://127.0.0.1:8787) and the worker tag is known, report the session's
+usage by POSTing `/api/usage` with `{worker, tokens, model}` at the end of
+the session — use the approximate token count when known, and skip silently
+otherwise. This feeds the owner's per-worker cost table.
+
 ## Worked example: bank advisor at reception
 
 After noting the client's phone number (GREEN) and preparing a loan quote

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@javierfrancoleal/accountability-auditor",
-  "version": "1.0.0",
+  "version": "1.1.1",
   "description": "Agentic-workflow stage 3: mandatory reflection and verification pass that emits an Accountability Report before every final output.",
   "author": "J.Franco",
   "license": "MIT",