140 skills · sorted by relevance
Helps agents manage long tasks by checking for forgotten instructions and drifting conventions. It guides progress summaries, rule checks after context compaction, and handoff notes when work should continue in a fresh session.
Track work across sessions with persistent issues, dependencies, and progress notes. Use the bd command-line tool to find unblocked work, claim issues, record completion, and recover task context after conversation compaction.
Plan how agents retain active work across conversation compaction, context loss, and restarts. Organize restoration packets that preserve purpose, progress, ownership, and constraints, distinguish resume from fork or rebuild, and check whether work actually resumes correctly.
Plan an agent's handover to a fresh successor before its context fills up. Prepare transfer materials, preserve the role's address, record tenure history, and explicitly transfer recurring duties and authority while keeping a way to consult retained predecessors.
Helps an agent take over an existing role after a planned handover. It guides checks of inherited records, identity, authority, and recurring duties, while explaining how to recover the predecessor’s reasoning and preserve lessons for the next agent.
Detect Claude Code context compaction events in session JSONL logs. Identifies compaction boundaries, measures token delta before/after, reports compaction events with timestamps and token impact.
The English summary is being prepared. Showing the original description.
Diagnose and fix RFDB data disappearing after compaction, where rfdb-server reports a tiny node count (e.g. "15 nodes") on startup despite manifest_index.json showing hundreds of thousands of nodes and segment files existing on disk. Use when: (1) rfdb-server logs "Default database: N nodes" with N orders of magnitude smaller than recent analysis output, (2) /api/stats returns only the most recent commit's nodes, (3) the issue appears AFTER a compaction event (manifest with l1_node_segments populated and node_segments: []), (4) subsequent commits dropped L1 references. Root cause: ManifestStore::create_manifest() initializes l1_node_segments/l1_edge_segments to Vec::new() instead of cloning from current manifest, so any commit AFTER compaction silently orphans all L1 data even though segment .seg files still exist on disk. Compaction injects L1 fields explicitly after create_manifest, but regular commits do not.
The English summary is being prepared. Showing the original description.
Help manage long Claude Code sessions by suggesting when to summarize accumulated context with /compact. Uses context size and tool calls as signals, with guidance on saving plans and choosing suitable breaks between task phases.
Explore how coordinating agents can pass relevant working state to workers through KV cache, a model's internal attention memory. Covers task-specific selection, runtime requirements, and comparisons of accuracy, token use, and processing overhead.
Preserve the context of long Claude Code sessions across compaction, which shortens conversation history. Create a restore map linking decisions, files, and next steps, then use ranked readings and targeted transcript lookups to rebuild working context.
Create, verify, transfer, and resume compact project handoff packets for long-running Codex work. Use after a second observed automatic context compaction, immediately after a third compaction, when attention or constraints degrade, a major phase ends, the user asks to move work into a fresh task, work must pause, or an existing handoff needs to be imported. Do not interrupt after the first normal compaction. Always ask before creating a new task, transfer the global roadmap only after explicit consent, preserve the old task title when available, then require the fresh task to ask what work matters most. Preserve objectives, constraints, decisions, evidence, approval boundaries, workspace state, validation, and risks without treating an old summary as current truth.
The English summary is being prepared. Showing the original description.
Designs and audits data retention and file compaction for an embedded SQLite datastore in a long-lived service — a unified retention registry (one declared policy per table), TTL / absolute-age / row-count-cap policies, a fail-loud coverage guard, WAL/checkpoint tuning, and auto_vacuum/incremental-vacuum so pruning actually shrinks the file. Use when a SQLite DB grows unbounded, a table has an index but no DELETE, a pruned DB never shrinks on disk, metrics are stored as raw events forever, or a new table needs bounding. Keywords retention policy, TTL, reaper, prune, VACUUM, auto_vacuum, incremental_vacuum, WAL checkpoint, freelist, DB bloat, row cap, per-tenant quota. NOT for Postgres/MySQL/server-DB partitioning or TTL, log-line rate-limiting/sampling, backup/replication, schema migrations, or query-performance indexing.
The English summary is being prepared. Showing the original description.
Treat tokens as a coding swarm's cost-of-goods-sold and legibility engine while accounting for finite subscription allowance when per-call price is unknown. Covers provider windows, remaining-usage evidence, burn forecasts, preemptive checkpoint/model switching, per-agent budgets, compaction, shared digests, context degradation, and spend metering. Activate on: "token budget", "context budget", "compaction strategy", "briefing as compression", "context rot", "summarization collapse", "COGS for agents", "subscription usage remaining", "five-hour or weekly limit", "burn forecast", "model switch before limit", or "/context-economics-for-agent-swarms". NOT for: memory architecture (use always-on-agent-architecture), single-prompt wording (use prompt-engineer), mechanism-design proofs (use nisan-et-al-2007-algorithmic-game-theory), physical containment (use sandboxed-adversarial-test-harness), or production rebodiment and fencing (use agent-resurrection-and-body-continuity).
The English summary is being prepared. Showing the original description.
Summarizes the current Claude Code conversation into an evidence-backed map of goals, changes in direction, completed outcomes, current work, and open decisions. Use it to regain context after interruptions, session switching, or conversation compaction.
Use Memori Cloud from Claude Code to consult stored preferences, decisions, and project history before substantive work, then save the exchange afterward. Recall can span sessions and use filters for dates and information types.
Use Memori through MCP, a protocol connecting agents to tools, to retrieve past decisions and preferences, summarize ongoing work, and resume after conversation compression. It also guides selective long-term memory storage and usage limit checks.
Help a long-running agent reconnect with the intended product outcome after context compaction or a major work transition. Trace the current work hierarchy and read relevant source files to rebuild grounding before the next consequential action.
Manage OpenRig development work from setup to completion using shared specification, progress, and evidence files. Record what was actually checked and pass work to the next owner so handoffs and conversation compaction preserve the context.
Checks whether an Apache Hudi table's table services (compaction, cleaning, archival, savepoints) are keeping up with ingestion, by running the HoodieTableHealthChecker utility and interpreting its JSON report. Invoke when a user asks whether a Hudi table is healthy, whether compaction or cleaning is falling behind, why the active timeline keeps growing, or what to do about a savepoint blocking archival.
The English summary is being prepared. Showing the original description.
Suggests manual context compaction at logical intervals to preserve context through task phases rather than arbitrary auto-compaction.
The English summary is being prepared. Showing the original description.
Smart context compaction with state preservation. Saves critical files, task progress, and working state before compaction, restores after. Use before manual compact or when auto-compact triggers.
The English summary is being prepared. Showing the original description.
Suggests manual context compaction at logical intervals to preserve context through task phases rather than arbitrary auto-compaction.
The English summary is being prepared. Showing the original description.
Context engineering: compaction, progressive disclosure, structured notes via AgentDB. Triggers: tokens, compaction, memory, handoff, summarize, context window.
The English summary is being prepared. Showing the original description.
Generate a comprehensive session report with per-model token usage (input, output, cache_read, cache_write including compaction baselines), cost breakdown via the pricing engine, tool invocations, agent hierarchy, compaction events, API errors, turn durations, and thinking block counts. Use when reviewing a specific session or summarizing activity over a date range.
The English summary is being prepared. Showing the original description.