🧠 Mnemosyne — an error-memory plugin for DeepSeek Harness

"Don't step on the same rake twice. And if you do — at least make sure it doesn't happen a third time."

Mnemosyne is a plugin for DeepSeek Harness that remembers your agent's mistakes and refuses to let it repeat them. It learns from your own screw-ups: catches an error → remembers it → comes up with a fix → simply won't let it happen again.

Why you want this

Agents are people too — they trip over the same things. Wrong path, forgotten argument, a drive that doesn't exist. Mnemosyne watches over this and turns every stumble into a lesson: it records the error, writes the remedy and — the important part — suggests the actual fix, so the agent can just keep going.

This is not "yet another logger". This is a guard dog that actually steps in: a repeated failing call simply never executes.

How it works

1. Catches. The plugin listens to tool results (tools/result) and agent errors (agent/error). A call fails — it already knows.
2. Remembers. Every error becomes a signature (tool :: category :: message) with a repeat counter.
3. Figures out the fix. On the second occurrence an automatic remedy is written: file errors → "check the path", schema errors → "validate arguments", and so on. If an LLM is configured, the plugin asks the model for a precise fix.
4. Suggests the correction. Saying "check the path" is lazy. Mnemosyne says what exactly to fix. It spots a missing M:\ drive? It offers D:\ and shows you the corrected command, ready to copy.
5. Blocks the repeat. Same tool, same arguments — the call is intercepted before dispatch. The agent gets an explanation and a fix instead of another error.
6. Learns from success. Once the agent finally nails it, the lesson flips to fixed ("the remedy worked").

Feature list

Feature	What it does
🎯 Auto-learning	catches, remembers, classifies (fs / schema / parse / timeout / guard / runtime)
🩹 Auto-remedy	a recipe for every repeat + a concrete fix (a ready-to-use command)
🛡️ Guard	a repeated failing call is blocked before it runs
⏱️ Time window	only blocks recent repeats (5 minutes by default)
🕊️ Amnesty	after 3 blocks the guard backs off — it won't choke a tool forever
🤖 LLM remedies	the model writes the fix (when a provider is configured), otherwise a template
💾 Persistence	lessons survive Harness restarts
🧹 Rules	/mnemosyne forbid <tool> / allow — manual bans
📊 Reports	/mnemosyne status and /mnemosyne report — stats and top lessons
🖥️ Live panel	a card in the Run view: counters, categories, recent blocks

Install

Drop src/host.js and src/client.js into your project and wire them in via an overlay:

# cordis.yml
- insert:
    - id: mnemosyne
      name: './src/host.js'
      config: {}

pnpm dsh web --patch ./cordis.yml

The panel shows up in the Run card and the /mnemosyne command becomes available in chat.

Commands

/mnemosyne status              — list lessons (error + remedy + fix)
/mnemosyne report              — summary: categories, top lessons, blocks, timeline
/mnemosyne forbid <tool>       — ban a tool entirely
/mnemosyne allow <tool>        — lift the ban
/mnemosyne clear               — wipe the memory

The model also gets mnemosyne_status and mnemosyne_learn tools.

Live example

$ Get-Content "M:\autofix\data.txt"
  → Cannot find drive. A drive with the name 'M' does not exist.

$ Get-Content "M:\autofix\data.txt"        # again
  → Mnemosyne remembered the lesson: recurring + remedy

$ Get-Content "M:\autofix\data.txt"        # third time
  → ⛔ Mnemosyne: blocked — the same failing call to 'pwsh' recurs (2x).
    Suggested fix: Replace M:\ with D:\ in the command — e.g. Get-Content "D:\autofix\data.txt"

The agent copies the fix, runs it successfully — and the lesson closes as fixed.

Under the hood

• Host (src/host.js) — the brain: lesson registry, classifier, autofix, guard, persistence, command, RPC.
• Client (src/client.js) — a live browser panel (React, no JSX).
• Plain JavaScript, no build step — a stock Cordis plugin for Harness.

License

MIT — take it, tweak it, share it. If you build something cool on top of it, tell us!

---

Made for DeepSeek Harness. Questions, ideas, bugs — Issues are welcome.