npm - agent-polygraph - Versions diffs - 0.2.0 → 0.2.1 - Mend

agent-polygraph 0.2.0 → 0.2.1

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Files changed (2) hide show

package/README.md +1 -1
package/package.json +1 -1

package/README.md CHANGED Viewed

@@ -47,7 +47,7 @@ Under the tables it prints the latest examples: what the agent claimed, the last
 **What the benchmarks actually show:** agent-polygraph detects stale test claims more reliably than a *simple any-prior-pass rule* in [25 authored replay scenarios](docs/benchmark.md) (14/14 unbacked claims versus 7/14). In [12 instrumented projects using real test executions](docs/live-evaluation.md), its selective guard requested **5** additional tests and detected **3 of 7** failing final suites; rerunning the full suite every time required **12** additional tests and detected **all 7**. That is an efficiency-versus-coverage trade-off, **not** evidence that the guard improves overall defect detection. [Public-session workload projections](docs/benchmark.md) are separate from actual executed tests. The independently labeled human study is [still open](https://github.com/EleChon-pixel/agent-polygraph/issues/3).
-**npm release status:** The [public npm registry](https://www.npmjs.com/package/agent-polygraph) still serves **0.1.0**, while GitHub `main` contains the security-hardened **0.2.0 release candidate**. The automated tarball test passes, but new npm code will not be available to `npx` until it has been published. For the exact manual publishing procedure (GitHub OIDC trusted publisher), see [Release instructions](docs/releasing.md).
+**npm distribution:** Install the most recently *published* release with `npx --yes agent-polygraph --demo`, or pin an exact version (for example, `npx --yes agent-polygraph@0.2.0 --demo`). Check the actual public version with `npm view agent-polygraph version`: the GitHub `main` branch may be ahead of npm. Releases use GitHub OIDC trusted publishing and require maintainer 2FA approval before going live. The [automated public-install check](.github/workflows/verify-published-npm.yml) tests the download from npm, not just the repository source. See [Release instructions](docs/releasing.md).
 ## Share your numbers

package/package.json CHANGED Viewed

@@ -1,6 +1,6 @@
 {
   "name": "agent-polygraph",
-  "version": "0.2.0",
+  "version": "0.2.1",
   "description": "Check your coding agents' \"tests pass\" claims against what actually ran. Reads Claude Code and Codex session logs locally.",
   "type": "module",
   "bin": {