← Back to Skills Marketplace
omaression

Auto Improving Agent

by omaression · GitHub ↗ · v1.0.0-alpha · MIT-0
cross-platform ✓ Security Clean
167
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install auto-improving-agent
Description
Automatically capture corrections, failures, and reusable discoveries into `.learnings/` files using signal-based filtering. Triggers when the user corrects...
README (SKILL.md)

Self-Improving Agent

Capture what matters. Ignore noise. Promote proven patterns. Automate all of it.

Source of truth

  • .learnings/LEARNINGS.md — corrections, env configs, reusable fixes, architecture decisions
  • .learnings/ERRORS.md — tool/command failures with fixes
  • .learnings/FEATURE_REQUESTS.md — missing capabilities worth tracking
  • .learnings/ARCHIVE.md — entries scored out during retention sweeps (never injected into context, but searchable)

Write gate

Before logging anything, the candidate must pass at least ONE filter:

Filter Weight Description
Correction ALWAYS Omar explicitly corrected the agent
Recurrence HIGH Same issue hit 2+ times (check existing entries)
Cost-to-rediscover HIGH Would take >2 tool calls to figure out again
Blast radius MEDIUM Affects multiple skills, projects, or workflows
Decay risk MEDIUM Non-obvious env/config detail that changes rarely

If NONE match → do not log. This replaces any arbitrary line-count threshold.

Never log:

  • routine successes
  • facts obvious from docs or code
  • one-off tasks with no recurrence potential
  • anything already in MEMORY.md, SOUL.md, USER.md, or AGENTS.md

Entry format

LEARNINGS.md:

- [YYYY-MM-DD] [Category]: [Actionable takeaway]

Categories: Correction, Env, Workflow, Testing, Skills, Git, Architecture

ERRORS.md:

- [YYYY-MM-DD] [Tool]: [What failed] → [Fix]

Mark fixed items with [fixed]. Delete stale entries during retention sweeps.

Retention gate

Instead of a hard line cap, score each entry periodically:

Signal Score
Referenced or applied in last 30 days +3
Matches active project context +2
Direct correction from Omar +2
Has prevented a repeat error +3
Env/config still valid +1
Superseded by newer entry −5
>90 days old, never referenced −3

Action:

  • score ≥ 2 → keep
  • 0 ≤ score \x3C 2 → archive to .learnings/ARCHIVE.md
  • score \x3C 0 → delete

Run this sweep during heartbeat maintenance (every ~3 days) or when LEARNINGS.md feels noisy.

Automated triggers

These fire without user prompting:

  1. Post-task scan: After multi-step tasks, check for retried commands, error→workaround sequences, or avoidable file reads. If found, evaluate against write gate and log if it passes.

  2. Session-start sweep: On .learnings/LEARNINGS.md read, flag entries >90 days old for retention scoring.

  3. Promotion detector: After logging, scan for entries with the same [Category] tag appearing 3+ times. If found, auto-suggest a one-liner promotion to:

    • behavior/style → SOUL.md
    • workflow/process → AGENTS.md
    • tool/env gotcha → TOOLS.md

  4. Cross-session pattern detection: When memory_search returns a daily note describing a workaround, check if .learnings/ already has it. If not and it passes the write gate, log it.

Dedup

Before logging, scan existing entries for near-duplicates. If the lesson already exists, only update it if the new version is sharper or more general.

Quality bar

Every entry must help a future session avoid wasted work in under one glance.

Usage Guidance
This package appears to be what it says: a local helper that suggests and scores learnings and adds a bootstrap reminder. A few practical notes before enabling: - The SKILL.md promises automated background triggers (post-task scans, periodic sweeps). The code included only provides a bootstrap reminder and a local retention_scorer.py utility; automatic capture would rely on the agent following the instructions or additional hook implementations. If you expect fully automatic logging, verify how your agent/platform will implement those triggers. - The skill will read and write files under .learnings/ and may modify LEARNINGS.md when the scorer is run (retention_scorer.py). Run the scorer with --dry-run first to see planned changes before letting it write files. - There are no network calls or credential requests in the provided files, but the instructions do tell the agent to consult workspace files and memory_search. If your workspace contains secrets or sensitive files, decide whether you want the agent to index or copy that content into .learnings/ (the skill's write gate discourages logging obvious facts or single-use items, but that's an instruction — not an enforced safeguard). - Review the retention_scorer.py and the hook files yourself (they are small) and run the included tests locally if you want assurance about behavior. If you accept that the agent will be permitted to read workspace files and write to .learnings/, this skill is coherent and low-risk. If you require true autonomous capturing of failures without additional platform support, request or implement the missing event handlers instead of relying on the SKILL.md promises.
Capability Analysis
Type: OpenClaw Skill Name: auto-improving-agent Version: 1.0.0-alpha The skill bundle provides a structured framework for an OpenClaw agent to manage its own 'learnings' and error logs within a `.learnings/` directory. It includes a Python maintenance script (`scripts/retention_scorer.py`) for log rotation and TypeScript/JavaScript hooks to inject behavioral reminders during the agent's bootstrap process. The logic is transparent, localized to the workspace, and lacks any indicators of data exfiltration, malicious execution, or harmful prompt injection.
Capability Assessment
ℹ Purpose & Capability
The README/SKILL.md describes automated, signal-driven logging (post-task scans, session sweeps, promotion detector). The provided hook code only injects a virtual SELF_IMPROVEMENT_REMINDER.md at agent bootstrap and the included Python script is a local retention scorer. There is no code here that actually implements background 'post-task' automatic capture or an autonomous logger; that behavior would have to be implemented by the agent following the SKILL.md instructions or by additional platform hooks. This is a mild incoherence between the claimed autonomous triggers and the code present, but not malicious.
✓ Instruction Scope
SKILL.md instructs scanning and writing files under .learnings/, performing retention scoring, and using memory_search to detect cross-session patterns. All of these actions are within the stated purpose (capturing learnings) and the code supports the retention scoring and bootstrap reminder. The instructions do ask the agent to read workspace files and memory_search results (expected for this feature) but do not direct data to external endpoints or request unrelated system credentials.
✓ Install Mechanism
No install spec and no external downloads — the skill is instruction/code-only. The included hook and Python script are local files (no network fetch or archive extraction). This is low-risk from an installation perspective.
✓ Credentials
The skill requests no environment variables, no credentials, and no config paths. Its operations are limited to reading/writing .learnings/ files and using platform memory_search; that scope is proportional to the stated purpose.
✓ Persistence & Privilege
always:false (default), and the hook only injects a virtual reminder at agent bootstrap. The skill does not request permanent platform-wide privileges, does not modify other skills' configs, and does not enable itself automatically across all agents.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install auto-improving-agent
  3. After installation, invoke the skill by name or use /auto-improving-agent
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0-alpha
Initial release of self-improving-agent — automates capturing actionable corrections, errors, and discoveries to `.learnings/` files. - Logs only high-value learnings, errors, and requests using a targeted signal-based filter. - Periodically scores and prunes logged data to retain only proven, still-relevant entries. - Promotes recurring or transferable lessons to higher-level skill, agent, or tool docs. - Deduplicates and sharpens entries for high-impact, future-proof guidance. - Automates all triggers to minimize user intervention and ensure continuous improvement.
Metadata
Slug auto-improving-agent
Version 1.0.0-alpha
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is Auto Improving Agent?

Automatically capture corrections, failures, and reusable discoveries into `.learnings/` files using signal-based filtering. Triggers when the user corrects... It is an AI Agent Skill for Claude Code / OpenClaw, with 167 downloads so far.

How do I install Auto Improving Agent?

Run "/install auto-improving-agent" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Auto Improving Agent free?

Yes, Auto Improving Agent is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Auto Improving Agent support?

Auto Improving Agent is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Auto Improving Agent?

It is built and maintained by omaression (@omaression); the current version is v1.0.0-alpha.

More Skills
self-improving agent
Captures learnings, errors, and corrections to enable continuous improvement. Use when: (1) A command or operation fails unexpectedly, (2) User corrects Clau...
⬇ 463283 · by pskoett
Skill Vetter
Security-first skill vetting for AI agents. Use before installing any skill from ClawdHub, GitHub, or other sources. Checks for red flags, permission scope, and suspicious patterns.
⬇ 259410 · by spclaudehome
Polymarket
Query Polymarket prediction markets. Check odds, find trending markets, search events, track price movements.
⬇ 232503 · by joelchance
Self-Improving + Proactive Agent
Self-reflection + Self-criticism + Self-learning + Self-organizing memory. Agent evaluates its own work, catches mistakes, and improves permanently. Use when...
⬇ 200423 · by ivangdavila
Github
Interact with GitHub using the `gh` CLI. Use `gh issue`, `gh pr`, `gh run`, and `gh api` for issues, PRs, CI runs, and advanced queries.
⬇ 191113 · by steipete
ontology
Typed knowledge graph for structured agent memory and composable skills. Use when creating/querying entities (Person, Project, Task, Event, Document), linkin...
⬇ 190156 · by oswalpalash

💬 Comments