move-doctor 0.2.0-dev.24edc5d → 0.2.0-dev.d0feceb

package/README.md

@@ -10,27 +10,7 @@ Deterministic Sui Move scanner. Rules grounded in [The Move Book](https://move-b
 npx move-doctor@latest
 ```
 
-Scans the current directory and prints a health score, a per-bucket breakdown, and next steps. Add `--verbose` for file/line refs and fix hints. Sample output:
-
-```
-  ╭────────────────────────────────────────────────────────────────────────╮
-  │                                                                        │
-  │ 44 / 100   poor                                          movebook_gaps │
-  │ ███████████████████████████████░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░ │
-  │                                                                        │
-  │ 2 modules · edition 2024.beta · Sui 1.70.2 · scanned in 1.3s           │
-  │ 22 findings  >  4 errors · 3 warnings · 15 info                        │
-  ╰────────────────────────────────────────────────────────── move.doctor ─╯
-
-  security     >  4 errors · 3 warnings
-  conventions  >  5 info
-  testing      >  5 info
-  functions    >  2 info
-  idioms       >  2 info
-  macros       >  1 info
-
-  > Run --verbose for file refs and fix hints.
-```
+Scans the current directory and prints a 0–100 health score, a severity breakdown by area, and next steps. Add `--verbose` for file/line refs and fix hints. With the Sui CLI on `PATH`, its `sui move build --lint` pass runs alongside and feeds the same score.
 
 ## Install as an agent skill
 
@@ -40,20 +20,15 @@ npx move-doctor@latest install
 
 Writes `<cwd>/.claude/skills/move-doctor/SKILL.md`. Claude Code picks it up automatically — type `/movedoctor` or ask it to "run move doctor" and it'll scan, group findings, and fix them.
 
-## CI gate
+## CI
 
-Use `--score` for a numeric output and gate PRs on a budget:
+`npx move-doctor@latest install` can set up a GitHub Actions workflow that scores every push and pull request. To gate by hand, fail the build below a score budget:
 
 ```bash
 SCORE=$(npx move-doctor@latest . --score)
 [ "$SCORE" -ge 80 ] || exit 1
 ```
 
-## Requirements
-
-- Node.js ≥ 20
-- (Optional) Sui CLI on PATH — when present, `sui move build --lint` runs in parallel and its 6 built-in lints feed the same score.
-
 ## Source
 
 - Source code: [github.com/notmatical/move-doctor](https://github.com/notmatical/move-doctor)

package/dist/cli.js

@@ -3886,7 +3886,7 @@ var renderWorkspaceJson = (workspace, result) => JSON.stringify(
 var renderWorkspaceScoreOnly = (result) => String(result.aggregateScore.score);
 
 // src/cli/utils/branded-header.ts
-var VERSION = "0.2.0-dev.24edc5d";
+var VERSION = "0.2.0-dev.d0feceb";
 var printBrandedHeader = () => {
   const wordmark = highlighter.accent(hyperlink("move.doctor", HOMEPAGE));
   process.stdout.write(
@@ -4042,14 +4042,9 @@ curl --fail --silent --show-error \\
   https://move.doctor/prompts/move-doctor-agent.md
 \`\`\`
 
-It's a scan \u2192 group \u2192 triage \u2192 fix \u2192 re-score loop that edits the working tree directly (never commits, never opens PRs). For each finding, fetch the matching per-rule fix recipe on demand \u2014 \`<bucket>/<rule>\` comes straight from the diagnostic's rule id:
+It's a scan \u2192 triage \u2192 fix \u2192 re-score loop that edits the working tree directly (never commits, never opens PRs). Every finding in \`--json\` carries its own \`fixHint\` and \`citation\` \u2014 fix straight from those; there's nothing else to fetch.
 
-\`\`\`bash
-curl --fail --silent --show-error \\
-  https://move.doctor/prompts/rules/<bucket>/<rule>.md
-\`\`\`
-
-If the fetch fails (offline / site down), fall back to: run \`move-doctor --verbose\`, fix errors first (\`security/*\` and \`abilities/*\` findings are real vulnerabilities, not style), apply each finding's \`fixHint\` verbatim, and re-run until the score stops rising. Never suppress a rule unless you can explain why the surrounding code is a documented exception.
+If the fetch fails (offline / site down), fall back to: run \`move-doctor --verbose\`, fix errors first (\`security/*\` and \`abilities/*\` findings are real vulnerabilities, not style), apply each finding's \`fixHint\`, and re-run until the score stops rising. Never silence a finding unless you can explain why the surrounding code is a documented exception.
 
 ## Command
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "move-doctor",
-  "version": "0.2.0-dev.24edc5d",
+  "version": "0.2.0-dev.d0feceb",
   "description": "A deterministic linter for Sui Move. Rules cited via the Move Book, with a SKILL.md for coding agents.",
   "keywords": [
     "sui",