@wipcomputer/wip-ldm-os 0.4.85-alpha.19 → 0.4.85-alpha.20

package/README.md

@@ -22,7 +22,18 @@ Open your AI and paste this:
 ```
 Read https://wip.computer/install/wip-ldm-os.txt
 
-Check if LDM OS is already installed. If it is, run ldm install --dry-run and show me what I have and what's new.
+Use the install document and live local checks as the source of truth. Do not search memory or prior notes for this install.
+
+Check if LDM OS is installed (`which ldm && ldm --version`). Branch.
+
+If installed: run `ldm status` and show me the version table and what's new for each component. Then ask:
+- Do you have questions?
+- Want to see a dry run?
+
+If yes to dry run, run `ldm install --dry-run`.
+
+If I say install, run:
+`npm install -g @wipcomputer/wip-ldm-os@latest && ldm install && ldm doctor`
 
 If not, walk me through setup and explain:
 
@@ -35,7 +46,14 @@ Then ask:
 - Do you have questions?
 - Want to see a dry run?
 
-If I say yes: Install the CLI first (npm install -g @wipcomputer/wip-ldm-os) and then run ldm install --dry-run.
+If yes to dry run, install the CLI first:
+`npm install -g @wipcomputer/wip-ldm-os`
+
+Then run:
+`ldm init --dry-run`
+
+If I say install, run:
+`ldm init`
 
 Show me exactly what will change. Don't install anything until I say "install".
 ```

package/SKILL.md

@@ -27,6 +27,12 @@ metadata:
 
 # LDM OS
 
+## Source of truth
+
+Use this install document and live local checks as the source of truth. Do not search memory or prior notes for this install.
+
+Do not run GitHub release commands during the install-state flow. Do not run `gh release list` or `gh release view` unless the user explicitly asks for release notes. Install-state answers should come from local commands, `ldm status`, and npm package metadata.
+
 ## Step 1: Check if installed
 
 ```bash
@@ -62,12 +68,9 @@ Always a table. Every component with an update gets its own row.
 
 **Do NOT skip this step.**
 
-```bash
-gh release list --repo wipcomputer/<repo-name> --limit 5 --json tagName,name --jq '.[]'
-gh release view <tag> --repo wipcomputer/<repo-name> --json body --jq .body
-```
+Use the output of `ldm status`, installed package metadata, and npm metadata. Do not use GitHub release commands here.
 
-Translate release notes to user language. Every bullet answers "what changed for ME?"
+Translate available update information to user language. Every bullet answers "what changed for ME?" If the status output does not include enough detail for a component, say that clearly and do not invent release notes.
 
 Good: "Your AIs now explain what LDM OS actually does when you ask them to install it"
 Bad: "Restored rich product content to SKILL.md"

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@wipcomputer/wip-ldm-os",
-  "version": "0.4.85-alpha.19",
+  "version": "0.4.85-alpha.20",
   "type": "module",
   "description": "LDM OS: identity, memory, and sovereignty infrastructure for AI agents",
   "engines": {
@@ -18,9 +18,10 @@
   "scripts": {
     "build:bridge": "cd src/bridge && npm install && npx tsup core.ts mcp-server.ts cli.ts openclaw.ts --format esm --dts --clean --outDir ../../dist/bridge",
     "build": "npm run build:bridge",
-    "prepublishOnly": "npm run build:bridge && npm run validate:bin-manifest",
+    "prepublishOnly": "npm run build:bridge && npm run validate:bin-manifest && npm run test:install-prompt-policy",
     "validate:bin-manifest": "node scripts/validate-bin-manifest.mjs",
     "test:skill-frontmatter": "node scripts/test-skill-frontmatter.mjs",
+    "test:install-prompt-policy": "node scripts/test-install-prompt-policy.mjs",
     "test:installer-update-tracks": "node scripts/test-installer-update-tracks.mjs",
     "test:installer-hook-toolname": "node scripts/test-installer-hook-toolname.mjs",
     "test:installer-target-self-update": "node scripts/test-installer-target-self-update.mjs",

package/scripts/test-install-prompt-policy.mjs

@@ -0,0 +1,60 @@
+import { readFileSync } from "node:fs";
+import { join } from "node:path";
+import { fileURLToPath } from "node:url";
+
+const repoRoot = fileURLToPath(new URL("..", import.meta.url));
+
+const files = [
+  "README.md",
+  "SKILL.md",
+  "shared/templates/install-prompt.md",
+];
+
+const contents = Object.fromEntries(
+  files.map((file) => [file, readFileSync(join(repoRoot, file), "utf8")]),
+);
+
+const failures = [];
+
+for (const file of ["README.md", "shared/templates/install-prompt.md"]) {
+  const text = contents[file];
+  for (const phrase of [
+    "Use the install document and live local checks as the source of truth. Do not search memory or prior notes for this install.",
+    "If installed: run `ldm status`",
+    "If yes to dry run, run `ldm install --dry-run`.",
+    "`npm install -g @wipcomputer/wip-ldm-os@latest && ldm install && ldm doctor`",
+    "Then run:\n`ldm init --dry-run`",
+    "If I say install, run:\n`ldm init`",
+  ]) {
+    if (!text.includes(phrase)) {
+      failures.push(`${file} missing install prompt phrase: ${phrase}`);
+    }
+  }
+  if (text.includes("If it is, run ldm install --dry-run")) {
+    failures.push(`${file} still tells installed users to start with ldm install --dry-run`);
+  }
+}
+
+const skill = contents["SKILL.md"];
+for (const phrase of [
+  "Do not run GitHub release commands during the install-state flow.",
+  "Do not run `gh release list` or `gh release view` unless the user explicitly asks for release notes.",
+  "Use the output of `ldm status`, installed package metadata, and npm metadata.",
+  "Do not use GitHub release commands here.",
+]) {
+  if (!skill.includes(phrase)) {
+    failures.push(`SKILL.md missing install policy phrase: ${phrase}`);
+  }
+}
+
+if (/gh release (list|view) --repo/.test(skill)) {
+  failures.push("SKILL.md still includes concrete gh release commands in the install flow");
+}
+
+if (failures.length > 0) {
+  console.error("install prompt policy checks failed:");
+  for (const failure of failures) console.error(`  - ${failure}`);
+  process.exit(1);
+}
+
+console.log("install-prompt-policy: LDM OS prompt and install doc agree");

package/shared/templates/install-prompt.md

@@ -6,7 +6,18 @@ Open your AI and paste this:
 
 Read https://wip.computer/install/wip-ldm-os.txt
 
-Check if LDM OS is already installed. If it is, run ldm install --dry-run and show me what I have and what's new.
+Use the install document and live local checks as the source of truth. Do not search memory or prior notes for this install.
+
+Check if LDM OS is installed (`which ldm && ldm --version`). Branch.
+
+If installed: run `ldm status` and show me the version table and what's new for each component. Then ask:
+- Do you have questions?
+- Want to see a dry run?
+
+If yes to dry run, run `ldm install --dry-run`.
+
+If I say install, run:
+`npm install -g @wipcomputer/wip-ldm-os@latest && ldm install && ldm doctor`
 
 If not, walk me through setup and explain:
 
@@ -19,7 +30,14 @@ Then ask:
 - Do you have questions?
 - Want to see a dry run?
 
-If I say yes: Install the CLI first (npm install -g @wipcomputer/wip-ldm-os) and then run ldm install --dry-run.
+If yes to dry run, install the CLI first:
+`npm install -g @wipcomputer/wip-ldm-os`
+
+Then run:
+`ldm init --dry-run`
+
+If I say install, run:
+`ldm init`
 
 Show me exactly what will change. Don't install anything until I say "install".