- Home
- Skills
- Agents & Orchestration
- forge refine
Works with the AI Agents you already use
forge refine
Automated lifecycle management for agent skill libraries using evidence-based retention and consolidation rules.
$7
forge refine
Example session with this skill installed
A new skill 'api-auth-handler' in the 'security' domain just passed the gate with a VERIFIED tier. Record the promotion and check if the library is still within size limits.
- Read your context and instructions
- Compiled the forge refine
I have promoted the api-auth-handler skill to the library and updated the ledger state. A snapshot of the previous version (if any) has been stored in the version history.
Skill Promotion and Library Audit
| Action | Entity | Status | Details |
|---|---|---|---|
| Promotion | api-auth-handler | SUCCESS | Moved to library/ from candidates/ |
| Tier Assignment | security domain | VERIFIED | Initial retention counter set to 3 |
| Versioning | _versions/api-auth-handler/ | CREATED | Snapshot of current state archived |
| Compile Check | Library Size | PASS | Total active skills: 14 / 20 cap |
| Domain Density | security | OPTIMAL | 2 active skills (below consolidation threshold) |
Next steps
- Monitor the next 3 calibration runs to observe the "hit" rate for the new
api-auth-handler. - If another security skill is created, trigger a merge via
skill-creatorto maintain consolidation discipline. - Run
python scripts/compile_check.pybefore the next deployment to ensure word-count compliance.
Connects securely to your tools. The creator never sees your data.
About this skill
Maintaining a high-performing agent skill library requires strict discipline. This skill manages the lifecycle of your library by promoting successful candidates, tracking hit/miss ratios, and enforcing size constraints to prevent performance degradation.
What it does
- Ledger management tracks which skills are active, their domain, and their current performance tier.
- Automated promotion moves gated candidates into production folders while snapshotting prior versions for safety.
- Evidence-based retention implements a hit/miss counter that automatically archives underperforming skills before they propagate errors.
- Library consolidation flags redundant skills in the same domain to force merges, optimizing for the highest possible performance lift.
- Constraint enforcement runs periodic checks to ensure word counts and library size caps are never exceeded.
How it works
- Promote candidates by running the ledger script with the gate verdict tier to move files from staging to the library.
- Record outcomes after every task attempt to update the retention counter based on success or failure.
- Audit the library using compilation checks to detect violations of size or word-count discipline.
- Trigger merges when the system identifies sibling skills in the same domain, handing them back to the creator for consolidation.
Frameworks & tools
The skill interfaces with local file systems and Python-based CLI tools (ledger.py, compile_check.py). It follows the SkillFlow architecture for automated skill evolution and archival.
Why this beats prompting it yourself
Manual library management leads to "skill bloat" where too many overlapping prompts confuse the agent. This skill uses a mathematical leash—based on the SkillFlow arXiv:2604.17308 findings—to prune failures and merge redundancies automatically, ensuring your agent's context remains lean and effective.
Use cases
- Lifecycle automation for multi-agent systems that generate and test their own tools.
- A/B testing prompts by tracking real-world "hit" rates and archiving losers.
- Version control for agent capabilities by maintaining a structured snapshot history.
- Performance optimization by preventing library accumulation from exceeding architectural limits.
Known limitations
This skill does not write or edit prompt content itself. It requires the existence of ledger.py and compile_check.py scripts to perform file operations and audits.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Agents & Orchestration.
Does not write or edit skill content. Requires local Python scripts for ledger and compilation checks. Lifecycle transitions are evidence-driven, not manual.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 6 days ago