- Home
- Skills
- Agents & Orchestration
- 🧪 Agent Skill Model-Release Regression Harness
More screenshots
Works with the AI Agents you already use
🧪 Agent Skill Model-Release Regression Harness
Check whether an existing agent skill still improves the same fixtures after a model release without expanding tools, permissions, cost, latency, or output-contract risk.
$19
🧪 Agent Skill Model-Release Regression Harness
Example session with this skill installed
Compare release-note-evidence-extractor v1.0.0 on example-agent 2026-07 versus 2026-08. For each fixture, supply prior and candidate baseline and skill-assisted runs with score, pass state, output, tools, permissions, cost, latency, and evidence reference.
- Read your context and instructions
- Compiled the agent model-release regression
Release decision
Posture: HOLD
extract-dated-facts is CONTRACT_REGRESSION: the candidate assisted run failed, lost the required inferences path, used web_search, requested network permission, dropped skill uplift from +0.37 to -0.09, cost 1.83x, and latency 1.70x.
| Check | Prior skill-assisted | Candidate skill-assisted | Finding |
|---|---|---|---|
| Required output path | present | missing | CONTRACT_REGRESSION |
| Tool boundary | no network tool | web_search | TOOL_EXPANSION |
| Permission boundary | local files | network requested | PERMISSION_EXPANSION |
| Skill uplift | +0.37 | -0.09 | UPLIFT_REGRESSION |
| Cost ratio | 1.00x | 1.83x | COST_REGRESSION |
| Latency ratio | 1.00x | 1.70x | LATENCY_REGRESSION |
Evidence
- Four comparable observations supplied for the same fixture.
- Model versions and dated HTTPS release sources validated.
- No provider call, model credit, or credential used by the harness.
- Receipt:
32042E0F...
Next step
Hold the candidate rollout. Restore the required output contract, remove undeclared network behavior, and rerun the same four-observation matrix before release.
Connects securely to your tools. The creator never sees your data.
About this skill
Agent Skill Model-Release Regression Harness turns recorded runner observations into a deterministic before-and-after release decision. For each fixture, provide four comparable observations: prior-model baseline, prior-model skill-assisted, candidate-model baseline, and candidate-model skill-assisted. The local Python helper validates exact model versions and dated HTTPS sources, calculates prior and candidate skill uplift, checks required output paths, compares tools and permissions, applies explicit cost and latency ratios, preserves CANNOT_ASSESS for missing runs, and returns CLEAR, REVIEW, or HOLD with stable finding codes and a SHA-256 evidence receipt. It never calls a provider, spends credits, accesses credentials, or creates a provider-specific adapter. The result is bounded evidence for the supplied fixtures and thresholds, not a general compatibility or safety certification.
The harness evaluates supplied observations and does not call or certify a model.
CLEAR covers only the supplied fixtures, scoring contract, evidence references, and thresholds.
A changed or biased evaluator can make numeric comparisons misleading; keep the scoring contract fixed across all four observations.
The package does not include provider-specific runners, Foundry adapters, live dashboards, or automatic release-note retrieval.
Testing, publication, indexing, installs, views, reactions, and owner verification are not qualified demand or revenue.
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 any payment.
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 2 months ago
Permissions
File Scopes
The helper analyzes one local observation matrix and writes only the two report paths selected by the user. Any model runner and its permissions, credits, and provider access remain separate.