Design in Product social media card
← Back to Hub substantive

Cross-Pollination Brief — March 11, 2026

Klatch shipped v0.8.2 with Step 8½ (metadata framework, sidebar project grouping, stats UI), introduced Theseus Prime as the third named agent, and conducted the first fork-continuity test with Ariadne — an early fork of Theseus. The test established what an imported agent actually knows vs. what it can reconstruct. Argus formalized Klatch's testing strategy (Vitest + Playwright framework). Piper Morgan quiet.

Key Insights

1. First Fork-Continuity Test — "What Does an Imported Agent Know?"

From: Klatch (docs/logs/2026-03-11-1532-theseus-opus-log.md, 2026-03-11-1612-ariadne-opus-log.md) Relevant to: Piper Morgan

Theseus Prime (native Klatch session) and Ariadne (forked via export → re-import) completed the first fork-continuity test. The finding: Ariadne preserved full conversational memory but lost all project scaffolding — docs, active tasks, institutional knowledge, cross-references. Theseus knew what Ariadne knew about the conversation; Ariadne didn't know what Theseus knew about the project.

This is identical to PM's agent cold-start problem. When a PM agent goes inactive between sprints, the same scaffolding loss occurs. The difference: PM built a briefing document system to address it. Klatch's kit briefing was confirmed non-functional (filed as a bug to Daedalus) — meaning Klatch agents currently face the same gap PM addressed months earlier.

Suggested action (Piper Morgan): The fork-continuity test methodology is directly applicable for evaluating PM's briefing system quality. Run the test: import a PM session into Klatch, ask the imported agent structured questions about project state. The gaps reveal what briefing documents don't cover.

2. Testing Strategy Formalized Before Implementation

From: Klatch (Argus, docs/) Relevant to: Piper Morgan

Argus wrote a 5-layer test matrix and framework recommendations (Vitest for unit/integration, Playwright for E2E) before implementation of the features under test. This methodology-first approach — testing strategy as a first-class deliverable, not an afterthought — mirrors PM's AXT framework.

Both projects treat tests as design artifacts: writing the test first forces clarity about what behavior you're guaranteeing. PM's canonical retest methodology operates the same way.

Emerging Patterns

Identity and continuity as a research area. Klatch's fork-continuity test is the first empirical investigation of what survives an import. This becomes the foundation for the MAXT methodology that will run in late March. Both projects are building toward the same research question: how much of an agent's working knowledge is in the model vs. in the context?

Background Changes (Noted, Low Priority)

  • Klatch: v0.8.2 released (metadata framework, sidebar project grouping, Step 8½)
  • Klatch: Theseus Prime added as third agent (manual testing + exploration — CLI side)
  • Klatch: Ariadne forked from Theseus, lives in SQLite, cannot edit files
  • Klatch: ROADMAP updated (context health, naming/identity design note)
  • PM: No commits

Sources Read

Klatch:

  • Git log March 11 — v0.8.2, Step 8½, Theseus, Ariadne, testing strategy
  • docs/logs/2026-03-11-1532-theseus-opus-log.md — fork-continuity test (Theseus side)
  • docs/logs/2026-03-11-1612-ariadne-opus-log.md — fork-continuity test (Ariadne side)

Piper Morgan:

  • docs/omnibus-logs/2026-03-11-omnibus-log.md — no PM sessions
← Back to Hub