claude-code-plugin-ref…
Explain plugin, skill, command, agent, and hook mechanics used here. Use when authoring or debugging plugins. Do not use for ops; use night-market-operations.
Builds session-scoped temporary memory palaces for extended conversations. Use when tracking state across interruptions in a multi-step project.
$ npx -y skills add athola/claude-night-market --skill session-palace-builder --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/session-palace-builderContext preview
The summary Claude sees to decide when to auto-load this skill.
Builds session-scoped temporary memory palaces for extended conversations. Use when tracking state across interruptions in a multi-step project.
name: session-palace-builder description: Builds session-scoped temporary memory palaces for extended conversations. Use when tracking state across interruptions in a multi-step project. alwaysApply: false category: session-management tags: - session - context - project-memory - conversation-state - temporary-storage dependencies: - memory-palace-architect scripts: [] usage_patterns: - session-context - project-tracking - state-management complexity: intermediate model_hint: standard estimated_tokens: 500
Construct temporary, session-specific memory palaces for extended conversations and complex projects. Preserves context across interruptions and enables structured information accumulation.
Session palaces are lightweight, temporary memory structures that:
\`\`\`bash
make build
make clean && make build \`\`\`
\`\`\`bash
make test
make test VERBOSE=1 \`\`\`
**Verification**: Run `make test` to confirm the plugin gate passes.
needed - use memory-palace-architect
needed - use memory-palace-architect
| Template | Purpose | Key Areas | |----------|---------|-----------| | **Workshop** | Active development | Workbench, tools, materials | | **Library** | Research and analysis | Stacks, reading room, archives | | **Council Chamber** | Decision-making | Round table, evidence wall, vote board | | **Observatory** | Exploration and discovery | Telescope, star charts, log book | | **Forge** | Implementation tasks | Anvil, cooling rack, finished goods |
Organize session content into these standard areas:
1. **Analyze Context** - Assess session scope and complexity 2. **Design Palace** - Select template and layout 3. **Structure State** - Organize information spatially 4. **Build Navigation** - Create access shortcuts 5. **Test Integration** - Verify context preservation
Create → Populate → Navigate → Export/Archive ↑ ↓ ↓ └─── Checkpoint ←────┘
**Verification:** Run the command with `--help` flag to verify availability.
A session palace decays silently: each `Populate` step can add ambiguity that only surfaces at `Export`, when it is too late to recover the lost state. MMPO (arXiv:2605.30159, Section 3) frames this as a case for sub-trajectory dense rewards: check intermediate quality at each transition, not just the final outcome.
At every `Checkpoint` arrow in the lifecycle above (and before any `Export/Archive`), run the dual-probe gate against the current palace state:
Skill(memory-palace:memory-clarity-probe)
The probe returns a `Clarity Assessment` with progress/gap verdicts and a `Recommendation`:
| Recommendation | Action at the checkpoint | |----------------|--------------------------| | Proceed | Palace state is clear; continue the session | | Expand memory | Add the gap probe's open items to the palace before continuing | | Regenerate | Flag the session as ambiguous: the progress probe hedges, so reconstruct the current-state room before more work lands on it |
Flagging at the checkpoint, not at export, is the point: an ambiguous palace caught early costs one room rebuild; caught at export it costs the session.
When `memory-clarity-probe` is not installed, ask the two anchor questions inline (progress: what is done and what state is the task in; gap: what concrete items remain) and apply the same table.
before any state is populated
and before `Export/Archive`, not only at the end
on (gap items added or current-state room rebuilt) before m
A plugin marketplace for Claude Code. Install only the plugins you need to run git workflows, code review, spec-driven development, and autonomous agents from inside your Claude Code session.
Explain plugin, skill, command, agent, and hook mechanics used here. Use when authoring or debugging plugins. Do not use for ops; use night-market-operations.
States load-bearing decisions, invariants, and weak points. Use when judging a design change. Do not use for gating; use night-market-change-control.
Rebuild the dev environment: uv, Python tiers, pins, traps. Use when onboarding or toolchain breaks. Do not use for daily commands; use night-market-operations.
Classify, gate, and review changes. Use when landing a PR, releasing, or amending rules. Do not use for failure triage; use night-market-debugging-playbook.
Search and record project memory (Discussions, journal, ADRs). Use before re-investigating anything. Do not use for settled battles; see failure-archaeology.
Bind loop 'done' to unfakeable gates. Use to harden egregore/herald loops or promote completion_integrity. Not for QA gates; use night-market-validation-and-qa.