@aimlsuperagent/agent 0.1.5 → 0.1.6

package/AGENTS.md

@@ -18,11 +18,15 @@ The framework must be:
 1. Read `REPO_SOURCE_OF_TRUTH.json` before making structural claims about this repo.
 2. Read `WORKING_NOTES.md` only when present and directly relevant.
 3. Prefer targeted search over loading broad folders.
-4. Do not add secrets, tokens, private URLs, account IDs, or credential values to examples.
-5. Keep examples generic unless explicitly marked as fictional.
-6. Make small diffs with a clear reason.
-7. Run the fastest meaningful verification after changes.
-8. Update durable notes only when a fact will help future work.
+4. Before changing code, confirm which backend, service, deployment, or environment is live when relevant.
+5. Check deployment/proof notes when the task depends on release or live behavior.
+6. Inspect the relevant source file before proposing or applying changes.
+7. Treat old notes as suspect until verified against source or production reality.
+8. Do not add secrets, tokens, private URLs, account IDs, or credential values to examples.
+9. Keep examples generic unless explicitly marked as fictional.
+10. Make small diffs with a clear reason.
+11. Run the fastest meaningful verification after changes.
+12. Update durable notes only when a fact will help future work.
 
 ## Context Minimizer Rules
 
@@ -83,4 +87,3 @@ For release changes:
 - confirm repo status
 - confirm private/public state before publication
 - confirm no secrets are present
-

package/README.md

@@ -131,7 +131,7 @@ After `init` and `check`, tell your AI coding assistant:
 ```text
 Read AGENTS.md, REPO_SOURCE_OF_TRUTH.json, and WORKING_NOTES.md first.
 Use them as the project operating context.
-Before editing code, verify the source of truth, search only the files relevant to this task, make the smallest safe diff, run the fastest meaningful check, and update durable notes only if reality changed.
+Before changing code, confirm which backend, service, deployment, or environment is live when relevant; check DEPLOYMENT_LOG.md and PRODUCTION_CHECK.md when available; inspect the relevant source file; avoid stale notes; make the smallest safe diff; run the fastest meaningful proof; and update durable memory only if reality changed.
 Do not store secrets, credential values, private customer data, local machine paths, or scratch-only notes in committed files.
 ```
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@aimlsuperagent/agent",
-  "version": "0.1.5",
+  "version": "0.1.6",
   "description": "A token-efficient operating framework for AI coding assistants.",
   "type": "module",
   "bin": {

package/scripts/postinstall.js

@@ -14,6 +14,7 @@ Next steps:
 Then tell your AI assistant:
   Read AGENTS.md, REPO_SOURCE_OF_TRUTH.json, and WORKING_NOTES.md first.
   Use them as project operating context before editing code.
+  Before changing code, confirm which backend, service, deployment, or environment is live when relevant; check DEPLOYMENT_LOG.md and PRODUCTION_CHECK.md when available; inspect the relevant source file; avoid stale notes; make the smallest safe diff; run the fastest meaningful proof; and update durable memory only if reality changed.
 
 Docs:
   https://github.com/marvinbfreedman/aimlsuperagent

package/templates/AGENTS.template.md

@@ -13,6 +13,10 @@ Describe the project in one paragraph.
 ## Working Rules
 
 - Use targeted search before loading broad folders.
+- Before changing code, confirm which backend, service, deployment, or environment is live when relevant.
+- Check `DEPLOYMENT_LOG.md` and `PRODUCTION_CHECK.md` when the task depends on release or live behavior.
+- Inspect the relevant source file before proposing or applying changes.
+- Treat old notes as suspect until verified against source or production reality.
 - Make small, task-traceable diffs.
 - Verify production reality before changing code when the task depends on live state.
 - Do not store secrets in notes, examples, commits, or logs.