vic/vix/ledger
Claude Code and agent skill by vic/vix
- community
Checking that the SKILL.md file is still in the repository.
Knowledge-preservation law (ALL projects) — command ledger (recall before re-run), past-generation transcript mining, telemetry duty on every dispatch, archival. Use before ANY expensive command, before solving anything previously-solved-smelling, at every handoff.
Install
npx vskill@latest install vic/vix/ledgerSKILL.md
LEDGER LAW (vic, global) CVMN — knowledge is constructed; re-running/re-thinking the known = #2 burn after fat context.
STACK: /vic-stack — vic-recall for commands, vic-mine for past transcripts, quipu label-prefix scan for verified facts; save verified results to quipu.
COMMAND LEDGER (automatic — global cmd-ledger hook logs EVERY Bash cmd + FULL output)
~/.cache/claude-ledger/<project-key>/bash-YYYYMMDD.jsonl (ts+sid+cmd+out60K+err). Outside repos ⟹ no VCS churn.
- RECALL-BEFORE-RUN: b4 any expensive cmd (build/lint/suite/bench):
vic-recall '<pattern>'. Tree unchanged since entry ⟹ that output IS current; re-running = pure waste. - KILLS TAIL TRAP: tail/truncate in-context (correct), full output in ledger — missed detail ⟹ vic-recall, NEVER re-run for a bigger tail.
GENERATION MINING
vic-mine '<pattern>' — searches THIS project's past session transcripts (prior generations ran it / solved it / killed it). Snippets + generation files; deep-dive = vic-reader on the .jsonl, NEVER inline (huge). Ranking: curated project docs (track DEAD/state) FIRST; mine for what curation missed. Periodic distill (vic-scribe): notable threads → tracks/playbooks; notebooklm-pack for archival to 2nd brain (direction only, never soundness).
TELEMETRY DUTY
After EVERY Agent/Workflow: vic-telemetry add <agent> <model> <tier> <secs> <tokens> <outcome> [note] (usage in tool results). state.md blocks + HANDOFF carry BURN: line. vic-telemetry report = routing feedback: tok/run high + ok on easy class ⟹ downgrade tier in /dispatch; fails high ⟹ tighten spec FIRST (spec cheaper than tier). /dispatch table = LIVING policy; scribe edits w/ receipt.
BURN RATIOS (route by this)
fable ≈2× opus ≈4× sonnet ≈8× haiku. haiku 4.5 · sonnet 4.6 · opus 4.8 · fable 5 (retires ~2026-06-17).
Security
- Security scanPassed, 90 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingmodules/vic/ai/vic/skills/ledger/SKILL.md
Popularity
- 7-day trend score
- 0
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/vic/vix/ledger)