klava

Optimize messages for desired outcome - fix English, kill red flags, simulate recipient

Periodic intake pipeline - reads new data, triages, acts, updates knowledge base

State + Log convention for sourced facts. Every deal note, person note, org note follows this shape. Use when migrating notes, writing new entity notes, or auditing for unsourced claims.

Use Vadim's Klava context stack, Obsidian vault, vadimgest, and autonomy rules.

Source-backed batch triage and checkpointing for grouped intake/backfill jobs.

Turn grouped source-intake batches into durable Obsidian writeback with provenance, routing, and checkpoint verification.

Process grouped calendar source records into entity notes or Inbox notes with provenance and checkpoint verification.

Process sourced intake batches into inbox/state/log notes and advance checkpoints safely.

Process grouped vadimgest-style intake batches into durable notes with exact checkpoint commits and provenance-preserving writeback.

Plan and run AI-agent dynamic workflows for complex tasks that benefit from explicit orchestration, goal mode, subagents or simulated work packets, approval gates, integration, verification, and reusable workflow artifacts. Use when the user invokes this skill, asks for a swarm, subagents, parallel agents, a dynamic workflow, a large migration or audit, multi-track research plus implementation, or Claude Code-style workflow orchestration.

Canonical response style rules: direct, grounded, low-noise, and action-first for interactions where the user asked for practical execution.

Free space on the codex-klava root disk by purging /tmp junk and moving large data dirs to the mounted volume with symlinks. Use when root disk > 85% or Vadim says "free space", "disk full", "move to mnt".

Class-level workflow for triaging vadimgest edge/local source batches into sourced inbox notes and exact checkpoint commits.

Grouped edge/local source intake backfill into sourced inbox notes and entity notes, with exact checkpoint commit verification.

Vadim's English learning system — catch his grammar mistakes and filler ("like", "basically", "actually") in chat, turn them into Anki flashcards, run spaced-repetition review, and export an .apkg deck. Use whenever Vadim writes English in chat, asks to "review"/"quiz"/"show cards", asks to "export" the deck, or wants to practice/improve English.

Batch triage and writeback for evidence-fed intake slices (GitHub, Gmail, Signal, Dayflow, etc.) into Obsidian inbox or entity notes, with provenance and checkpoint discipline.

Run a local OpenRouter Fusion style multi-model deliberation using available CLI agents such as claude -p and codex exec. Use for deep research, high-stakes synthesis, architecture decisions, code review second opinions, product comparisons, strategy memos, contradiction finding, or any task where independent model perspectives and a judged synthesis are worth extra cost and latency.

Triage grouped email/intake batches into compact inbox notes and canonical backfills, preserving source provenance and avoiding one-note-per-row noise.

Turn grouped vadimgest/intake source batches into durable Obsidian notes with provenance, deduplication, and checkpoint-safe writeback.

Class-level workflow for compacting grouped vadimgest/intake batches into durable Obsidian notes with Q1–Q4 routing, deduping, and checkpoint-safe backfill.

Backfill grouped source-intake batches into durable notes with per-channel compaction, sourced provenance, and checkpoint-safe writeback.

Compact grouped vadimgest intake batches by channel/thread, preserve source provenance, and route only durable facts to Inbox or entity notes.

Process pre-batched vadimgest edge/local intake slices into compact sourced Inbox notes, deduping grouped channel/thread evidence and committing only the exact checkpointed source lines; references/2026-07-01-heartbeat-edge-backfill.md and references/2026-07-03-heartbeat-edge-backfill-batch.md

Run-class workflow for operating Hermes in autonomous mode: managing cron jobs, approval gates, and proposal/memory queues with safe verification.

How to choose, configure, and verify Hermes memory providers for reliability, privacy, and retrieval quality.

Deploy, verify, and operate the Hermes WebUI (nesquena/hermes-webui) as a service on codex-klava. Use when asked to set up, check, restart, or troubleshoot the web/phone UI for the Hermes agent.

Class-level workflow for handling Hermes skill/memory write-approval safety gates and safely clearing/processing pending writes.

Class-level workflow for draining batched edge/intake records into Obsidian inbox notes, deduping repeated signals, and safely committing checkpoints.

Grouped source-intake backfill workflow for compacting vadimgest/intake batches into durable notes and then running exact batch-specific checkpoint commits.

Class-level workflow for turning grouped vadimgest / intake batches into durable Obsidian writebacks, with compact thread-level triage, deduping, and checkpoint-safe commits.

Class-level workflow for minimizing burn on PB-scale storage while preserving reliability and processing throughput.

Route sparse intake batches that only expose headers, titles, or create/update metadata into durable inbox notes without over-inventing content.

Process mixed vadimgest-style intake batches into sourced inbox notes or entity writebacks, then advance the checkpoint safely.

Route vadimgest-style multi-source intake batches into compact Obsidian notes, inbox summaries, and checkpoint updates.

Safely update paginated markdown notes during source-backed intake/backfill work without corrupting content or losing checkpoint integrity.

Backfill grouped source batches into existing paginated notes safely, without duplicating sections or skipping checkpoint commits.

Backfill paginated vadimgest-style source batches into durable notes with deduped group compaction, source-provenance, and checkpoint verification.

Handle high-signal user requests with minimal friction, preserving directness, action-first execution, and explicit handling of recurring frustration signals.

Compact grouped source-intake batches into durable Inbox, State, and Log notes with full provenance.

Append grouped source-backed intake batches into existing Inbox notes without duplicating topics, source URIs, or checkpoint state.