@vheins/local-memory-mcp 0.14.10 → 0.15.1

package/dist/prompts/task-memory-executor.md

@@ -11,21 +11,22 @@ version: "1.5.4"
 tags: [workflow, task-execution, memory]
 ---
 
-## FSM — Main Loop
+## Main Loop
 
-Entry=S0 → S1 → G0 → S2 → S3 → S4 → S5 → S6 → S7 → S8 Exit=exhausted|blocked
+Entry=S0 → S1 → G0 → S2 → S3 → S4 → S5 → S6 → S7 → S8 → S9 Exit=exhausted|blocked
 Guard: S(N) req S(N-1)✅; dependency-ready filter (depends_on+parent_id done)
 
 S0 | sync: resolve identity (arg→auto `<runner>-<randomName>`, 1x reuse all loop) + task-list(ONCE) + handoff-list(pending, close stale) + audit stale in_progress(>30m, hydrate via task-detail) | — | filtered queue | —
 S1 | hydrate: task-detail ONCE per task — MUST cache, MUST reuse all steps, NO re-fetch | S0✅ | full task | —
 G0 | readiness: depends_on✅ AND parent_id✅? if all blocked → report blockers + pause | S1✅ | → S2 / skip+pick next | —
 S2 | claim: task-claim(with identity metadata) + task-update→in_progress(agent, role, identity) | G0✅ | ownership | —
-S3 | research: memory-search + standard-search(MANDATORY per task — even sub-agents/decomposed) + hydrate relevant | S2✅ | context | —
+S3 | research: check task-detail for suggested_skills — load each via skill() tool; then memory-search + standard-search(MANDATORY per task — even sub-agents/decomposed) + hydrate relevant | S2✅ | context | —
 S4 | execute: trace logic+callsites+docs — DO NOT infer from file presence; decompose if too broad | S3✅ | changes | —
 S5 | validate: tests + linters + type-check + browser(if UI — MANDATORY: console errors, overflow, responsive, core interactions) + logic audit all paths | S4✅ | verification | —
 S6 | finalize: task-update→completed(evidence: inspected files, verified logic, test results) + memory-store(insights) + standard-store(rules) + handoff(if work remains — with identity) + retrospective + report | S5✅ | completion | —
 S7 | commit: `type(scope): msg — {{task_title}} {{summary_task}} {{keyword}} #N` (fix|closes|resolve, extract N from metadata/URL) | S6✅ | git commit | —
 S8 | repeat → S0 | queue not empty | next task | —
+S9 | verify: confirm commit format compliance, task updated completed, no stale handoffs remain | S8✅ | verified | —
 
 ## Design Note
 
@@ -37,15 +38,16 @@ Main loop is intentionally infinite — runs until MCP task queue is fully exhau
 - **Max 2** parallel sub-agents; each executes EXACTLY 1 task
 - Fallback: sequential (1 concurrent) if no sub-agent capability
 
-## FSM — Blocker Handling
+## Blocker Handling
 
-Entry=S0 → S1 → G1|G2 → S2 Exit=unblocked|deferred
+Entry=S0 → S1 → G1|G2 → S2 → S3 Exit=unblocked|deferred
 
 S0 | detect: task-update→blocked with reason | is blocked? | blocker comment | —
 S1 | classify: regex-match comment against patterns below | S0✅ | internal solvable | external | —
 G1 | internal solvable? | S1✅ | → S2 auto-create | —
 G2 | external? (awaiting user, API down) | S1✅→external | → keep blocked, no auto task | —
 S2 | create fix task: code=`{parent}-FIX-{unix}`, title=`FIX: [{parent_title}] — Resolve: {reason}`, priority=4(HIGH), phase=blocker-resolution, parent=current_id, depends_on=parent, tags=[blocker-fix,auto-generated], metadata={triggered_by, blocker_reason, timestamp, identity} | G1✅ | blocker fix task | —
+S3 | verify: confirm fix task created, parent task dependency linked | S2✅ | verified | —
 
 ### Blocker Regex Patterns
 
@@ -56,10 +58,11 @@ S2 | create fix task: code=`{parent}-FIX-{unix}`, title=`FIX: [{parent_title}]
 | Config    | `/(config\|configuration\|setup\|env\|environment)\s+(missing\|not\s+set\|invalid)/i`                    | `DATABASE_URL not set`                            |
 | Test fail | `/(test\|build\|compile\|type\s+check)\s+(failed\|error)/i`                                              | `Type error: Property 'user' not on 'Request'`    |
 
-## FSM — Backlog Maintenance
+## Backlog Maintenance
 
-Entry=S0 → S1 Exit=promoted
+Entry=S0 → S1 → S2 Exit=promoted
 Guard: active queue empty?
 
 S0 | list backlog: task-list(status=backlog) | active queue empty? | backlog tasks | —
 S1 | promote top 20 by priority(5→1) to pending via task-update | S0✅ | tasks→pending | —
+S2 | verify: confirm promoted task count, check priority ordering correct | S1✅ | verified | —

package/dist/prompts/tech-affinity-scout.md

@@ -11,14 +11,19 @@ version: "1.0.0"
 tags: [planning, patterns, memory, tech-affinity]
 ---
 
-## FSM
+## Tech Affinity Scout
 
-Entry=S0 → S1 → S2 → S3  Exit=scouted
+Entry=S0 → S1 → S2 → S3 → S4 Exit=scouted
 Guard: S(N) req S(N-1)✅
 
-S0 | search: memory-search (current_tags=[tags]) + standard-search (stack=[tags]) | tags provided? | pointer rows | —
+S0 | search: memory-search (current_tags=[tags]) + standard-search (stack=[tags]) + web_search (current practices for stack) | tags provided? | pointer rows + web results | —
 S1 | hydrate: memory-detail for relevant pointers | S0✅ | full entries | —
-S2 | filter: pattern + decision + coding standard entries from similar stacks | S1✅ | applicable knowledge | —
-S3 | adapt: explain adaptation to current project; separate memory-derived vs standards | S2✅ | adaptation guide | —
+S2 | filter: pattern + decision + coding standard + web_search entries from similar stacks | S1✅ | applicable knowledge | —
+S3 | adapt: explain adaptation to current project; separate memory-derived vs standards vs web_search | S2✅ | adaptation guide | —
+S4 | verify: confirm all relevant patterns evaluated, adaptation guidance is actionable | S3✅ | verified | —
 
 Tags: {{tags}}
+
+## Delegation
+
+Web search MUST be delegated to a coding subagent (general/explore). Main agent must NOT execute web_search directly.

package/dist/prompts/technical-planning.md

@@ -11,7 +11,7 @@ version: "1.0.0"
 tags: [workflow, technical-planning]
 ---
 
-## FSM
+## Technical Planning
 
 Entry=create-task → task-claim  Exit=task-update completed|blocked
 Guard: A(N) req all prev A✅
@@ -26,6 +26,7 @@ A5 | api-contract-design | G2✅ | API contract | design/api/
 G3 | approve API contracts | A5✅ | → proceed | —
 A6 | derive total_sprints from roadmap timeline+capacity → generate 1 sprint-planning task per sprint(N) + allocation audit + MCP task tree(parent/child + depends_on) | G3✅ | roadmap + sprints-1..N + audit + tasks | tasks/ + MCP
 GF | final approve — blocked if sprint-N.md missing or MCP task tree not created | A6✅ | → documentation | —
+V1 | verify: confirm all artifacts exist at documented output paths, gate approvals recorded | GF✅ | verified | —
 
 ## Optional
 

package/dist/prompts/tool-usage-guidelines.md

@@ -8,9 +8,9 @@ version: "1.0.0"
 tags: [workflow, tooling, memory, policy, mcp]
 ---
 
-## FSM — Memory Flow
+## Memory Flow
 
-Entry=S0 → S1 → S2 → S3 → S4   Exit=maintained
+Entry=S0 → S1 → S2 → S3 → S4 → S5   Exit=maintained
 Guard: S(N) req S(N-1)✅
 
 S0 | navigate: memory-recap (overview) / memory-search (targeted) | — | pointer rows | —
@@ -18,22 +18,25 @@ S1 | hydrate: memory-detail (id or code) before relying on content | S0✅ | ful
 S2 | store: memory-store (durable only, tech tags, human title, aux in metadata) | — | stored | —
 S3 | maintain: memory-update / supersedes for changes; avoid duplicates | — | updated | —
 S4 | acknowledge: memory-acknowledge (used|irrelevant|contradictory) after code gen | code generated? | feedback | —
+S5 | verify: confirm acknowledge called after code gen, no duplicate memories created | S4✅ | verified | —
 
-## FSM — Standards Flow
+## Standards Flow
 
-Entry=S0 → S1 → S2   Exit=done
+Entry=S0 → S1 → S2 → S3   Exit=done
 Guard: S(N) req S(N-1)✅
 
 S0 | search: standard-search mandatory before code edit/test/refactor/migrate (query, lang, stack, repo) | — | applicable | —
 S1 | apply precisely as implementation rules | S0✅ | compliant code | —
 S2 | store: standard-store (1 rule/entry, name, content, context, version, lang, stack, tags, scope) | — | new entry | —
+S3 | verify: confirm standard-search was called, scope and tags correct | S2✅ | verified | —
 
-## FSM — Handoff & Claim Flow
+## Handoff & Claim Flow
 
-Entry=S0 → S1 → S2 → S3   Exit=resolved
+Entry=S0 → S1 → S2 → S3 → S4   Exit=resolved
 Guard: S(N) req S(N-1)✅
 
 S0 | check: handoff-list (status, from/to agent) + claim-list | — | active state | —
 S1 | ensure: NO handoff for completion summaries — use task-update or memory | — | valid | —
 S2 | create: handoff-create (unfinished work only, summary + structured context) / task-claim (ownership) | S0✅ | pending | claimed | —
 S3 | close: handoff-update (accepted|rejected|expired) / claim-release (stale ownership) | consumed | stale | resolved | —
+S4 | verify: confirm no stale handoffs remain, claims reflect actual ownership | S3✅ | verified | —

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@vheins/local-memory-mcp",
-	"version": "0.14.10",
+	"version": "0.15.1",
 	"description": "MCP Local Memory Service for coding copilot agents",
 	"mcpName": "io.github.vheins/local-memory-mcp",
 	"type": "module",