@dzhechkov/harness-cli 0.3.250 → 0.3.251

package/.dz-manifest.json

@@ -9,7 +9,7 @@
       },
       {
         "path": "README.md",
-        "sha256": "1559995482381fd6ac146e6d229cb7d33476cdca8fad21fd361467dd698bd065"
+        "sha256": "4ffd489223826f51e8777ce825d60552365012de11d408ed9cf8c407357c0229"
       },
       {
         "path": "coverage/coverage-final.json",
@@ -73,7 +73,7 @@
       },
       {
         "path": "package.json",
-        "sha256": "4e22f0553a19181eea750db59bd3eec84fd880473ac523705a1a26197665e87e"
+        "sha256": "4926d5f944db1dc7361d0ebd80be28f99455133796ce10970df553eb57c198e8"
       },
       {
         "path": "src/bin.ts",
@@ -101,5 +101,5 @@
       }
     ]
   },
-  "signature": "d+++ms/ySpdJ2w5VE7F3b8IaZLWhIYcrvTRT1nPQmi2OHDdrQZ4YBHHW2KNxrauUdTXuUf/7tD//NQiGCQgQDw=="
+  "signature": "UAmZ0MzvL3RreqyakIVHBNTDP3o2/9atXMcT/OuxBjkrLo8YC48WFe7bYEP6R2TXH1p1+m8m2HRU4PzveHBdAQ=="
 }

package/README.md

@@ -1175,6 +1175,38 @@ The `--bto` flag generates eval templates compatible with `/bto-test`:
 
 After scaffolding, fill in the SKILL.md protocol and run `/bto-test .claude/skills/my-skill` to evaluate.
 
+### dz backlog — brain-backed idea inbox (smart backlog)
+
+Throw ideas in as they come; the backlog dedups them semantically against everything you already
+captured (via the same agentdb vector engine the Brain uses — no second store), scores them against
+your goals, and picks what to work on when you have free time.
+
+```bash
+# 1. Capture an idea — it is instantly deduped + aligned
+dz backlog add "add a compounding metric to strengthen dz recall" --effort 3
+#   → [new]  e3  align 0.72   (or: [duplicate of 268d3cb1 @ cosine 0.95] / [related: …])
+
+# 2. Define your goals once ("map + compass") — alignment is scored against these
+dz backlog goals            # opens/creates .dz/backlog/goals.json; --validate checks it
+
+# 3. When you have free time — spin the roulette (weighted: alignment^α · recency · 1/effort)
+dz backlog roulette --seed 7
+#   → 268d3cb1 — add a compounding metric …  align 0.72 · effort 3  (add --commit to start it)
+
+# 4. Work it: stage an idea2prd hand-off (writes the PRD-pipeline input scaffold)
+dz backlog enrich 268d3cb1  # → features/<slug>/ scaffold; run the idea2prd-manual skill on it
+
+# 5. Or push it to Jira via the configured adapter (jira-mcp | copilot-mcp | none)
+dz backlog jira 268d3cb1    # 'none' (default) writes an auditable .dz/backlog/jira-outbox/<id>.json
+
+# Housekeeping: cluster near-duplicates (dry-run by default; --apply prunes vectors too)
+dz backlog harmonize --apply
+```
+
+When to use: any time an idea strikes mid-work — capture beats context-switching. Without the
+agentdb memory backend installed the dedup honestly degrades to exact-text and alignment to 0
+(`dz setup --memory agentdb` enables the semantic tier).
+
 ### dz install — install skills from any npm package
 
 ```bash

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@dzhechkov/harness-cli",
-  "version": "0.3.250",
+  "version": "0.3.251",
   "description": "The dz CLI — install AI skills for Claude Code, Codex, OpenCode, Hermes, OpenClaude, GitHub Copilot. 35 commands, 13 presets, 6 platform targets.",
   "type": "module",
   "license": "MIT",

package/sbom.json

@@ -25,7 +25,7 @@
       "hashes": [
         {
           "alg": "SHA-256",
-          "content": "1559995482381fd6ac146e6d229cb7d33476cdca8fad21fd361467dd698bd065"
+          "content": "4ffd489223826f51e8777ce825d60552365012de11d408ed9cf8c407357c0229"
         }
       ]
     },
@@ -185,7 +185,7 @@
       "hashes": [
         {
           "alg": "SHA-256",
-          "content": "4e22f0553a19181eea750db59bd3eec84fd880473ac523705a1a26197665e87e"
+          "content": "4926d5f944db1dc7361d0ebd80be28f99455133796ce10970df553eb57c198e8"
         }
       ]
     },