v0.16.1

fix(prompts): increase default list limit to 50 so task-memory-executor appears on page 1

v0.16.0 — Software Engineering Analyst Roles

v0.16.0 — Software Engineering Analyst Roles

Added

• 6 Software Engineering Analyst Roles: New slash command prompt definitions for structured SDLC workflows:
  • Scrum Master — Sprint planning, retrospectives, backlog grooming, and impediment resolution.
  • Business Analyst — Stakeholder requirement extraction, user story mapping, and acceptance criteria writing.
  • System Analyst — Architecture design, database schema planning, and API contract design.
  • QA Analyst — Test scenario design (positive/negative/monkey/security), QA execution, and regression testing.
  • Data Analyst — Data modeling, query optimization, schema planning, and migration testing.
  • Security Analyst — Security triage, vulnerability assessment, penetration test planning, and threat modeling.
• Agent Labels: Server instructions now display an agent: label next to each prompt, enabling clients to show role badges.

Changed

• Prompt Pagination: Increased default listing limit to accommodate the growing prompt library.

v0.15.1 — Enriched Prompt Definitions with Skill Context

v0.15.1 — Enriched Prompt Definitions with Skill Context

✨ New Features

Enhanced Prompt Definitions (8 enriched)

Adopted expanded context from vibe-coding-premium skills-src into MCP prompt definitions:

Prompt	Added Context
csl-from-docs	Atomic Entry Rules (9 rules), Output Contract (tool vs JSON fallback), Refusal Rules with JSON error template
csl-scrapper	Atomic Entry Rules, Output Contract with sub-page task creation, Refusal Rules
security-triage	web_search for CVE/CVSS lookup + delegation rule (MUST delegate to coding subagent)
tech-affinity-scout	web_search for current practices + delegation rule
review-and-audit	FORBIDDEN: NON-EXECUTION + SELF-CHECK guard sections
review-and-post-issue	FORBIDDEN + OUTPUT: GITHUB ONLY + SELF-CHECK guard sections
sentinel-issue-resolver	SENTINEL Protocol identity, OUTPUT: AUTONOMOUS ACTION, Intelligence Gathering (GH CLI), Finalization & Commit format
task-management-guidelines	Navigation rules, Detail Tools reference, Workflow Rules (transition safety, claiming, handoff, archiving)
learning-retrospective	Directives section (`type: mistake
memory-agent-core	Memory Guardian Protocol, Core Rules (hybrid search 0.55 threshold, acknowledge, coordination state prohibition)

Refinements

• All prompt definitions now have a consistent S(N) verify step in their FSM
• Section headers updated for clarity and uniformity

📦 Assets

• npm: @vheins/local-memory-mcp@0.15.1
• Release artifacts attached below

v0.15.0 — Sequential Codes, Suggested Skills, Optional task_code

v0.15.0 — Sequential Codes, Suggested Skills, Optional task_code

Added

• Sequential Auto-Generated Codes: Task, memory, and standard codes are now auto-generated when omitted. Patterns: TASK-001, MEM-001, STD-001 (sequential per repository). No more random hard-to-read codes.
• Optional task_code: No longer required during task creation. Omit to get a sequential TASK-xxx code, or provide a custom code as before. Bulk items can also omit task_code.
• suggested_skills Field: Tasks can now carry a suggested_skills array of skill names. The task-memory-executor prompt reads this field and loads each skill via the skill() tool before execution.
• Cross-Reference Resolution: depends_on and parent_id now auto-resolve task codes within the same batch creation request, eliminating ordering constraints.

Changed

• Memory Codes: Replaced random 6-char codes (e.g., Q7PXYE) with sequential MEM-001 format.
• Standard Codes: Replaced random 6-char codes (e.g., J78C5E) with sequential STD-001 format.
• Prompt Tests: Updated assertions to match current FSM prompt formatting.
• Documentation: Updated tools-reference.md (EN/ID) and workflow prompts to reflect optional task_code, sequential codes, and suggested_skills.

Tests

283 tests passing, 34 test files, 0 failures.

v0.14.10

What's Changed

• Added git remote -v hint to prompt definitions for GitHub repo detection
  • export-task-to-github.md
  • import-github-issues.md
  • review-and-post-issue.md
  • sentinel-issue-resolver.md

v0.14.9

Changes

• refactor(prompts): convert server instructions.md to compact FSM format
  • Session start flow as FSM (S0→S1→S2)
  • Core workflows as compact data flows
  • Preserved all available prompts reference list

v0.14.8

Changes

• refactor(prompts): convert 25 prompt definitions to FSM format
  • Structured each prompt as state transitions (S0..SN, G0..GN)
  • Added explicit guards, outputs, and artifact paths per state
  • Restored missing template variables, version fields, and strict output format specifications
  • Fixed tool name ambiguity and parameter inference risks

v0.14.7

Changed

• Delete Confirmations: Replaced all native browser confirm()/alert() with SweetAlert2 modals across the dashboard UI. Delete confirmations now use styled, theme-aware dialogs.

v0.14.6

Fixed

• Dashboard Icons: Resolved 13 broken icon references across the Svelte UI. Added 10 missing SVG icon definitions to Icon.svelte. Replaced nonexistent icon names (x-circle, trash-2, edit-2, pencil, circle-alert, save, loader, check-check) with valid equivalents. svelte-check: 0 errors.

v0.14.5

Perubahan

🧪 Test & Type Check Fixes

• Fix spec_compliance test assertions for conditional structuredContent
• Fix standard.test to use structured: true in all store calls
• Fix e2e conflict text assertion (memory-delete → delete first)
• Fix tasks.bulk test assertions for new code-embedded summaries
• Fix tasks.e2e test timeout

🐛 Lainnya

• Fix a11y warning in FloatingChat.svelte (add keydown handler)
• Fix err type unknown in api.ts
• Exclude test dirs from root tsconfig