- Home
- Skills
- Testing & Debugging
- qa evidence freshness ledger
qa evidence freshness ledger
Verify if previous test results are still valid after code or environment changes to avoid redundant reruns.
Free
Works with the AI Agents you already use
qa evidence freshness ledger
Example session with this skill installed
Compare the current tree against the last test run in /reports/junit.xml. I updated the README and a CSS file in /assets, but the core logic in /src is unchanged. Can I skip the backend suite?
- Read your context and instructions
- Compiled the qa evidence freshness
I have analyzed the current tree against the evidence in /reports/junit.xml. The core logic remains identical, and the modified files are outside the defined backend coverage boundary.
QA Freshness Report: Backend Suite
| Attribute | Value |
|---|---|
| Check Name | backend-api-suite |
| Identity Match | MATCHED (src/ hashes identical) |
| Decision | FRESH |
| Last Run | 2023-10-24 14:22 UTC |
| Evidence Path | /reports/junit.xml |
| Coverage | 142/142 Assertions |
| External State | UNVERIFIED (Database fixtures not snapshotted) |
Identity Comparison
- Source (
/src): No changes detected (Hash:8f2a1b). - Dependencies (
package-lock.json): No changes detected. - Assets (
/assets): 1 file modified (ignored per backend scope). - Docs (
README.md): 1 file modified (ignored).
Decision Notes
The evidence is eligible for reuse because the backend logic and runner configuration are stable. However, because the database state was not snapshotted during the last run, this is marked as UNVERIFIED for external freshness. If the staging database has been migrated since the last run, a rerun is recommended.
Next steps
- Proceed with merge if database state is known to be stable.
- Authorize a
PARTIALrerun of integration tests if schema changes occurred. - Append this freshness decision to the PR summary.
Connects securely to your tools. The creator never sees your data.
About this skill
Software changes often outpace test cycles, leading to redundant reruns or risky assumptions about old results. This skill manages a QA Evidence Freshness Ledger to determine if existing test results still apply after code, dependency, or environment shifts.
What it does
- Verify evidence relevance by comparing current source hashes, lockfiles, and environment states against previous test records.
- Detect stale artifacts when source files, fixtures, or runtime configurations no longer match the latest recorded pass.
- Assess external state including live services, database snapshots, and expiry policies that invalidate local hashes.
- Generate rerun recommendations that target only the specific invalid tests to minimize CI/CD overhead.
- Document coverage gaps by identifying missing assertions, interrupted runs, or partial coverage in existing logs.
How it works
- Define the check by specifying the behavior, exact command, and required environment context.
- Boundary identification scans source files, dependencies, and safe configuration values to create a tree identity.
- Evidence inspection reads local artifacts to extract logs, timestamps, and previous runner outputs.
- Comparison and binding matches current tree state against the record to find identity or coverage drifts.
- Decision assignment outputs a status of
FRESH,STALE,PARTIAL, orFAILEDbased on the measured delta.
Frameworks & tools
This skill works with local file systems, version control systems like Git, and standard test runner outputs. It is compatible with any framework that emits readable artifacts, including Jest, Pytest, Cypress, and Playwright.
Why this beats prompting it yourself
General prompts often mistake a "passing" status for a "relevant" one, leading to false positives on dirty trees. This skill enforces a strict decision matrix that accounts for hidden invalidators like dependency lockfiles and external service expiry.
Use cases
- Selective test execution to skip expensive suites when non-relevant files are modified.
- Monorepo change analysis to identify which package results are still valid for a PR.
- Pre-deployment audits to confirm all required checks were actually run against the final commit.
Known limitations
This is not a tamper-proof attestation system or an executable cache. It cannot prove a test was meaningful or trustworthy, only that the inputs match the record.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Testing & Debugging.
Does not execute tests or install tools. Cannot verify the truthfulness of provided logs or prevent manual tampering of local artifacts. Requires readable runner output files to function.
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
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 13 days ago
- Free with an Agensi account