Skip to content

0.3.5: simpler code, same behavior

Latest
Latest

Choose a tag to compare

View all tags
@linxule linxule released this 06 Nov 22:37
v0.3.5
22a058f
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

This is a cleanup of the code the 0.3.x releases added, made before the next round of features. Outputs are unchanged: the old and new plugin gave the same results in 193 side-by-side scenarios on Claude Code, Codex, Copilot CLI and Kimi Code, with the one exception below.

Changed

  • If a project's config stores the Stage 1 counts as text ("5" rather than 5), the session-start and Stage 1 hooks now show 0, as /qual-status, the dashboard and the readiness check already did. The config schema allows only numbers there.
  • Each piece of shared logic now lives in one place: the Stage 1 progress formula and thresholds (stage1-thresholds.js), the analytical phase (phase.js) and the conversation log (conversation-log.js). Scripts that kept their own copies, or typed "10" and "10-15" by hand, now use these.
  • In the hook adapter, how a skill's Stop hook behaves on other platforms is now a stopHook field in its trigger table, and the schema tests read that table directly.
  • Rule updates check the phase without starting a separate process, so the end of each turn starts one fewer node process once a phase skill has run.

Upgrade

claude plugin marketplace update interpretive-orchestration
claude plugin update interpretive-orchestration@interpretive-orchestration

Then restart Claude Code. On other platforms, update the plugin the way you installed it. Codex asks you to trust the plugin's hooks again after an update.

Assets 2
Loading