Skip to content

Releases: nikolaylosev/sessionlens-vscode

0.1.109

Choose a tag to compare

@github-actions github-actions released this 29 Sep 14:23
4c0cc36

Fixed

  • A setting changed in settings.json (or VS Code Settings) is no longer switched back by a SessionLens page that
    saves something else before it has picked up the change. Only the values you change in the panel are written.

Changed

  • Gherkin findings carry their own source: in exported verdicts (verdicts.json) they are "gherkin" instead of
    "formal". The panel shows them as before.
  • New name: SessionLens: AI Agent Test Review in the stores and the Extensions view, and SessionLens in the
    sidebar (was "SessionLens for VSCode"). Its ID, nikolaylosev.sessionlens-vscode, is the same: nothing to
    reinstall, no data moves.
  • The extension's description in the Marketplace and Open VSX says what it finds, and its keywords now include
    Claude, Codex, Playwright, Cypress, test automation, CLAUDE.md and AGENTS.md.

Found a false finding? Please open an issue with a short, anonymized piece of the session: remove paths, user names, hosts, tokens and company names first. See Found a false finding?.

0.1.108

Choose a tag to compare

@nikolaylosev nikolaylosev released this 29 Sep 09:01
a366c3a

Fixed

  • Specification coverage: a requirement ID in the comment right above a test now links that test for the second and later tests in a file too. Before, in Java, Kotlin, C#, TypeScript/JavaScript, Go and Karate the comment was read as part of the test above it, so the test looked unlinked (test_without_requirement, and spec_uncovered for its requirement), and a short test above could take the ID.

Found a false finding? Please open an issue with a short, anonymized piece of the session: remove paths, user names, hosts, tokens and company names first. See Found a false finding?.

0.1.107

Choose a tag to compare

@nikolaylosev nikolaylosev released this 28 Sep 15:48
83a698a

Fixed

  • AI review through Claude Code or Codex CLI works when the path to the CLI contains non-Latin letters, for example a Windows user folder named in Cyrillic. Such a path used to be refused as "CLI not found".
  • On Windows the save dialog no longer suggests b.md for a name like a:b.md; it suggests sessionlens.md, as on macOS and Linux.

Changed

  • The readme asks you to report a false finding as an issue with an anonymized piece of the session. Quick start now says the Sessions list is below the panel.
  • THIRD-PARTY-NOTICES.md names every package inside vendor-eslint.js.

Found a false finding? Please open an issue with a short, anonymized piece of the session: remove paths, user names, hosts, tokens and company names first. See Found a false finding?.

sessionlens-vscode-0.1.106

Choose a tag to compare

@nikolaylosev nikolaylosev released this 28 Sep 12:47

First public release of SessionLens for VSCode.

SessionLens reviews Claude Code and Codex sessions in which an agent writes or fixes automated tests. It reads the
transcript, flags where the agent went wrong, lets you record a verdict on each finding, and turns the findings you
confirm into rules for CLAUDE.md, AGENTS.md or both.

Install

  • Marketplace: search for SessionLens in the Extensions view, or
    code --install-extension nikolaylosev.sessionlens-vscode
  • From this release: download sessionlens-vscode-0.1.106.vsix below, then in VS Code open Extensions →
    ⋯ → Install from VSIX…

Requires VS Code 1.85 or later.

What's in it

  • Imports Claude Code .jsonl, Codex rollout-*.jsonl, claude.ai conversations.json, /export and plain chat text.
  • Around 50 checks for process, methodology and test code (plan before code, tests actually run, weakened or weak
    assertions, sleeps and skips, fragile locators, API-specific mistakes, Gherkin, specification coverage).
  • Profiles for TypeScript/Playwright, Cypress, Detox, Python, Java/Kotlin, C#, API testing, Appium, Robot Framework
    and a generic one; the panel shows what each profile checks.
  • Static analysis: ESLint for TS/JS, tree-sitter for Java, C# and Python, a built-in parser for Robot Framework.
  • Optional semantic review by a model: Anthropic, OpenAI, Google, xAI, DeepSeek, Qwen, a local OpenAI-compatible
    server, or your Claude Code / Codex subscription through their CLIs. API keys are kept in VS Code's secret storage.
  • Calibration from your verdicts: checks that are often wrong on your sessions are demoted or switched off
    automatically; confirmed findings become proposed rules, a compressed policy or an Agent Skill.
  • Everything stays on your machine; see "Your data" in the README.

Notes

  • If you used an earlier local build (local.sessionlens-vscode), it is a separate extension: its data is not carried
    over. Export verdicts and rules there first if you need them, then uninstall it.
  • Full history: CHANGELOG.md