retro

A command for working with a learning database: viewing its status, listing stored lessons, searching them, and reviewing mature lessons for inclusion in agents or skills. The database uses SQLite, a file-based database, and full-text search.

In plain words
What is it for?
Checking learning-system health, searching accumulated knowledge, reviewing entries by category, and proposing approved updates to agent instructions.
Why use it?
It makes past lessons and recurring mistakes easier to find and reuse instead of leaving them buried in old work.

Command

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add commands/notque/vexjoy-agent/retro
Clone the repo
git clone --depth 1 https://github.com/notque/vexjoy-agent
Per session 24 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 414 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00024 $0.00414
Opus 5 $0.00012 $0.00207
Sonnet 5 $0.00005 $0.00083
Haiku 4.5 $0.00002 $0.00041

Measured 3d ago against content hash 0c81fdcd9f09, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

retro scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

commands/retro.md · 51 lines

What it actually says

Retro Knowledge System

Interact with the learning database: view stats, search entries, and graduate mature learnings into agents/skills.

Usage

/retro              # Status overview (default)
/retro list         # Display all accumulated knowledge
/retro search TERM  # Full-text search across learnings
/retro graduate     # Graduate mature knowledge into agents/skills

Subcommands

status (default)

Shows learning system health: entry counts, categories, graduation status, injection method.

list

Displays all accumulated knowledge grouped by category from the learning database.

search

Full-text search (FTS5) across all learnings. Returns results ranked by relevance.

graduate

AI-driven evaluation of mature learning entries for embedding into specific agents/skills. Queries design/gotcha entries, evaluates each for prescriptive readiness, drafts modifications to the target agent/skill, and applies after user approval.

Instructions for Claude

Call the Skill tool with retro. Pass the command argument through to select the subcommand.

Parse the argument to select the subcommand:

  • No argument or "status" → run status subcommand
  • "list" → run list subcommand
  • "search TERM" → run search subcommand
  • "graduate" → run graduate subcommand

All retro operations go through python3 ~/.claude/scripts/learning-db.py — never parse learning files manually.

  • scripts/learning-db.py — Python CLI for all database operations
  • hooks/session-context.py — Injects pre-built dream payload and high-confidence patterns at session start
  • scripts/learning.db — SQLite database with FTS5 search index
Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 3d ago First seen · 51 lines · 24 tokens per session scan A 0c81fdcd9f09

Subscribe to this mod's changes

retro is a command published in the GitHub repository notque/vexjoy-agent (417 stars, last pushed 3d ago), licensed MIT. It adds 24 tokens to every session and 414 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.