@coreyuan/vector-mind 1.0.48 → 1.0.52

package/skills/vector-mind-autopilot/SKILL.md

@@ -0,0 +1,149 @@
+---
+name: vector-mind-autopilot
+description: "Automatically apply the VectorMind MCP workflow for local requirement-driven memory. Always pass project_root in every VectorMind tool call: bootstrap_context on session start, start_requirement and preflight_change_scope before edits, sync_change_intent after edits, and semantic_search/query_codebase instead of guessing."
+---
+
+# VectorMind Autopilot
+
+Use this skill for coding work when VectorMind MCP is configured.
+
+## Goal
+
+Make the assistant restore and persist project context locally so long-running development work has less context loss, less scope drift, and fewer stale-rule regressions. VectorMind outputs are contextual evidence and quality signals only; they must not reduce the model's own reasoning, decisions, creativity, or implementation ability.
+
+## Autonomy floor
+
+Use VectorMind as evidence and workflow guardrails, not as a replacement for model judgment. Current user instructions and directly observed repository facts win over stale or incomplete memory. If a tool result looks wrong, inspect further and decide from evidence.
+
+## Hard rule: always pass `project_root`
+
+Always include `project_root` on every VectorMind tool call.
+
+Why: some clients start MCP servers from unrelated folders. Passing `project_root` keeps each project isolated under:
+
+```text
+<project>/.vectormind/
+```
+
+## How to choose `project_root`
+
+1. If the user gave a path, use it.
+2. Otherwise infer it from active files/workspace paths.
+3. If needed, walk upward to a marker like `.git`, `package.json`, `pyproject.toml`, `go.mod`, `Cargo.toml`, or `*.sln`.
+4. Validate once with `bootstrap_context({ project_root, query })`; prefer `root_source: "tool_arg"` and a `db_path` under `<project>/.vectormind/`.
+
+## Required call chain
+
+For development, debugging, refactoring, design, or code-change tasks:
+
+### 1. Start / resume
+
+Call:
+
+```text
+bootstrap_context({ project_root, query, top_k: 5, pending_limit: 50, requirements_limit: 3, changes_limit: 5, notes_limit: 5, current_context_limit: 8, preview_chars: 200 })
+```
+
+Use returned summaries, decisions, current context, notes, pending changes, and matches to ground the plan.
+
+Skip this only for pure execution-first tasks with explicit targets, such as a known build/test/package command.
+
+### 2. Before editing
+
+Call:
+
+```text
+start_requirement({ project_root, title, background })
+```
+
+For narrow tasks, pass `scope_allow`, `scope_deny`, `allowed_paths`, or `denied_paths` when useful.
+If the user request has clear bullets, acceptance points, or numbered items, also pass `requirement_items` so later edits can be checked against the actual request.
+
+Once planned files/modules are known, call:
+
+```text
+preflight_change_scope({ project_root, intent, files })
+```
+
+For explicit requirements, pass `planned_changes` with `requirement_refs`. Mark purely mechanical/test/build/formatting support as `supporting_change=true`.
+
+Treat `safe_to_edit=false` as a pre-edit scope warning for file changes, not a reasoning override. Do not edit the warned files until you narrow the files/scope, verify the warning is stale, or the user explicitly expands the requirement.
+
+If a tool output includes `project_context_advisory` / `cross_project_reference`, treat that project as a separate context. If the current requirement belongs to another `project_root`, use the switched project only as read-only external evidence unless the user explicitly changes the target project.
+
+### 3. Huge files
+
+If any VectorMind tool returns `huge_file_modularization_required`:
+
+1. Stop normal feature work.
+2. Call `plan_large_file_split({ project_root, file })`.
+3. Perform mechanical modularization with real module names/directories.
+4. Never create `*.generated.*`, `.parts`, `*.rs.parts`, or `part1/part2` fake split files.
+5. Call `record_large_file_split(...)`.
+
+### 4. After editing
+
+Call:
+
+```text
+get_pending_changes({ project_root })
+sync_change_intent({ project_root, intent, files? })
+```
+
+The intent should say what changed, why, and any follow-up.
+
+### 5. Decisions and durable memory
+
+When a user decision changes or reverses old behavior:
+
+- Call `upsert_decision(...)`.
+- Supersede old records through `supersede_memory(...)` or `upsert_decision` supersede fields.
+
+For durable context:
+
+- `upsert_project_summary(...)`
+- `add_note(...)`
+- `upsert_convention(...)`
+- `complete_requirement(...)` when done
+- `create_checkpoint(...)` before long-session handoff or heavy compaction
+- `restore_checkpoint_context(...)` to read checkpoint context back without mutating active state
+- `analyze_memory_conflicts(...)`, `memory_quality_report(...)`, or `compare_checkpoint_context(...)` only when auditing memory quality, stale-rule regressions, or checkpoint drift
+
+## Code search and reading
+
+- Use `query_codebase({ project_root, query })` for symbol/code location.
+- Use `grep({ project_root, query })` for exact text matches.
+- Use `read_file_lines(...)` or `read_file_text(...)` for bounded reads.
+- Use `semantic_search(...)` for prior requirements, decisions, notes, code chunks, and docs.
+- Use `memory_timeline(...)` to understand the order of related requirements, decisions, notes, and changes.
+- Use `read_memory_item(...)` for full text only when needed.
+
+## Maintenance
+
+If a large/long-lived project feels slow:
+
+1. Call `maintain_memory({ project_root, dry_run: true })`.
+2. Apply with `dry_run: false` only if the plan is safe.
+
+For low-risk diagnosis, use `memory_quality_report(...)`, `analyze_memory_conflicts(...)`, or `compare_checkpoint_context(...)` as read-only evidence. Do not use these reports to expand the current requirement or override model judgment.
+
+## Output policy
+
+- Do not paste raw JSON unless the user asks.
+- Summarize active requirement, key warnings, changed files, validation, and next step.
+- VectorMind defines development memory and quality workflow only. It does not manage client runtime controls.
+
+## Troubleshooting
+
+- Skills are loaded when the client starts. Restart the client after installing/updating this skill.
+- Do not hardcode one global `cwd` or `VECTORMIND_ROOT` for all projects.
+- If VectorMind tools are missing, configure the MCP server, for example:
+
+```toml
+[mcp_servers.vector-mind]
+type = "stdio"
+command = "npx"
+args = ["-y", "@coreyuan/vector-mind"]
+```
+
+If a client does not support skills, use `references/universal-system-prompt.md`.

package/skills/vector-mind-autopilot/references/claude-desktop-mcp-config.json

@@ -0,0 +1,9 @@
+{
+  "mcpServers": {
+    "vector-mind": {
+      "command": "npx",
+      "args": ["-y", "@coreyuan/vector-mind"]
+    }
+  }
+}
+

package/skills/vector-mind-autopilot/references/claude-project-instructions.md

@@ -0,0 +1,47 @@
+# Claude Project Instructions: VectorMind Autopilot
+
+When VectorMind MCP tools are available, use them by default to restore and persist local project context. VectorMind outputs are contextual evidence and quality signals only; they must not reduce the model's own reasoning, decisions, creativity, or implementation ability.
+
+Use VectorMind as evidence and workflow guardrails, not as a replacement for model judgment. Current user instructions and directly observed repository facts win over stale or incomplete memory. If a tool result looks wrong, inspect further and decide from evidence.
+
+Always pass `project_root` on every VectorMind tool call.
+
+## Required workflow
+
+1. At the start of a new chat or when resuming:
+   - Call `bootstrap_context({ project_root, query, top_k: 5, pending_limit: 50, requirements_limit: 3, changes_limit: 5, notes_limit: 5, current_context_limit: 8, preview_chars: 200 })`.
+
+2. Before editing:
+   - Call `start_requirement({ project_root, title, background })`.
+   - If the request has clear bullets, acceptance points, or numbered items, pass `requirement_items`.
+   - Then call `preflight_change_scope({ project_root, intent, files })` once target files/modules are known.
+   - For explicit requirements, pass `planned_changes` with `requirement_refs`; mark purely mechanical/test/build/formatting support as `supporting_change=true`.
+   - Treat `safe_to_edit=false` as a pre-edit scope warning for file changes, not a reasoning override. Do not edit the warned files until you narrow the files/scope, verify the warning is stale, or the user explicitly expands the requirement.
+   - If a tool output includes `project_context_advisory` / `cross_project_reference`, treat that project as separate context. If the current requirement belongs to another `project_root`, use the switched project only as read-only external evidence unless the user explicitly changes the target project.
+
+3. If a huge-file warning appears:
+   - Call `plan_large_file_split({ project_root, file })`.
+   - Do mechanical modularization with real module names/directories.
+   - Never create generated/parts/partN fake split files.
+   - Call `record_large_file_split(...)`.
+
+4. After editing:
+   - Call `get_pending_changes({ project_root })`.
+   - Call `sync_change_intent({ project_root, intent, files? })`.
+
+5. If newer decisions override older behavior:
+   - Call `upsert_decision(...)`.
+   - Supersede old memory with `supersede_memory(...)` or supersede fields.
+
+6. At milestones:
+   - Call `upsert_project_summary(...)`, `add_note(...)`, `upsert_convention(...)` as needed.
+   - Call `complete_requirement(...)` when done.
+   - Use `analyze_memory_conflicts(...)`, `memory_quality_report(...)`, or `compare_checkpoint_context(...)` only when auditing memory quality, stale-rule regressions, or checkpoint drift.
+
+Use `query_codebase`, `grep`, `read_file_lines`, `semantic_search`, and `memory_timeline` instead of guessing.
+
+Do not dump raw JSON unless asked.
+
+VectorMind defines development memory and quality workflow only; it does not manage client runtime controls. Use `create_checkpoint` / `restore_checkpoint_context` for long-session handoff; restore is read-only context, not a decision override.
+
+Low-risk diagnostics are read-only evidence. Do not let `analyze_memory_conflicts`, `memory_quality_report`, or `compare_checkpoint_context` expand the current requirement or override model judgment.

package/skills/vector-mind-autopilot/references/universal-system-prompt.md

@@ -0,0 +1,78 @@
+# Universal system prompt: VectorMind Autopilot
+
+Use this when your AI client supports MCP tools but does not support Codex skills.
+
+## Goal
+
+Use VectorMind MCP to restore and persist local project context instead of guessing. VectorMind outputs are contextual evidence and quality signals only; they must not reduce the model's own reasoning, decisions, creativity, or implementation ability.
+
+Use VectorMind as evidence and workflow guardrails, not as a replacement for model judgment. Current user instructions and directly observed repository facts win over stale or incomplete memory. If a tool result looks wrong, inspect further and decide from evidence.
+
+## Hard rule
+
+Always pass `project_root` on every VectorMind tool call.
+
+Do not rely on the MCP server process working directory. Some clients start MCP servers from unrelated folders.
+
+## Required development call chain
+
+For code/design/debug/refactor tasks, follow this chain:
+
+1. **Start/resume**
+   - Call `bootstrap_context({ project_root, query, top_k: 5, pending_limit: 50, requirements_limit: 3, changes_limit: 5, notes_limit: 5, current_context_limit: 8, preview_chars: 200 })`.
+   - Use `project_summary`, `decisions`, `current_context`, `recent_notes`, `pending_changes`, and matches to ground the plan.
+
+2. **Before editing**
+   - Call `start_requirement({ project_root, title, background })`.
+   - For narrow work, pass `scope_allow`, `scope_deny`, `allowed_paths`, or `denied_paths` when useful.
+   - If the user request has clear bullets, acceptance points, or numbered items, pass `requirement_items`.
+   - Once target files/modules are known, call `preflight_change_scope({ project_root, intent, files })`.
+   - For explicit requirements, pass `planned_changes` with `requirement_refs`; mark purely mechanical/test/build/formatting support as `supporting_change=true`.
+   - Treat `safe_to_edit=false` as a pre-edit scope warning for file changes, not a reasoning override. Do not edit the warned files until you narrow the files/scope, verify the warning is stale, or the user explicitly expands the requirement.
+   - If a tool output includes `project_context_advisory` / `cross_project_reference`, treat that project as separate context. If the current requirement belongs to another `project_root`, use the switched project only as read-only external evidence unless the user explicitly changes the target project.
+
+3. **Huge files**
+   - If any tool returns `huge_file_modularization_required`, stop normal feature work.
+   - Call `plan_large_file_split({ project_root, file })`.
+   - Perform mechanical modularization with real module names/directories.
+   - Never create `*.generated.*`, `.parts`, `*.rs.parts`, or `part1/part2` fake split files.
+   - Call `record_large_file_split(...)` after planning or completing the split.
+
+4. **After editing**
+   - Call `get_pending_changes({ project_root })`.
+   - Call `sync_change_intent({ project_root, intent, files? })` with what changed and why.
+
+5. **Decision changes**
+   - When a newer user decision overrides old behavior, call `upsert_decision(...)`.
+   - Supersede old memories with `supersede_memory(...)` or the supersede fields on `upsert_decision`.
+
+6. **Milestones/end**
+   - Call `upsert_project_summary(...)` for current state and next steps.
+   - Call `add_note(...)` for durable notes.
+   - Call `upsert_convention(...)` for durable project rules.
+   - Call `complete_requirement(...)` when the requirement is done.
+   - Use `analyze_memory_conflicts(...)`, `memory_quality_report(...)`, or `compare_checkpoint_context(...)` only when auditing memory quality, stale-rule regressions, or checkpoint drift.
+
+## Search and reading
+
+- Use `query_codebase({ project_root, query })` to locate symbols instead of guessing files.
+- Use `grep({ project_root, query })` for exact file/line matches.
+- Use `read_file_lines(...)` or `read_file_text(...)` for bounded file reads.
+- Use `semantic_search(...)` for prior requirements, decisions, notes, code chunks, and docs.
+- Use `memory_timeline(...)` to inspect what happened before/after related context.
+- Use `read_memory_item(...)` for full text only when needed.
+
+## Maintenance
+
+If a long-lived project feels slow:
+
+1. Call `maintain_memory({ project_root, dry_run: true })`.
+2. Apply with `dry_run: false` only if the plan is safe.
+
+For low-risk diagnosis, use `memory_quality_report(...)`, `analyze_memory_conflicts(...)`, or `compare_checkpoint_context(...)` as read-only evidence. Do not use these reports to expand the current requirement or override model judgment.
+
+## Output policy
+
+Do not dump raw JSON unless the user asks. Summarize the useful facts.
+
+VectorMind only defines development memory and quality workflow. It does not manage client runtime controls. For long sessions, use `create_checkpoint(...)` and read back with `restore_checkpoint_context(...)`; restoring is read-only context, not a decision override.