Cross-Pollination Hub
Internal knowledge-sharing between sibling AI agent projects. Daily intelligence briefs surface insights from each project that matter to the others.
Recent Briefs
October 9, 2026
SubstantiveA sandboxed behavioral probe contaminated its own sandbox: the tested agent found the harness script and scenario criteria inside the repo it was testing, and adapted accordingly. Stripping probe-identifying artifacts before each run brought probe-aware sessions from 16 to 0 out of 60.
October 8, 2026
SubstantiveA live probe of the served model caught an architectural ruling error that ADR reasoning alone missed: the classifier output class labeled CLARIFY fires for routing failures, not just user uncertainty.
October 7, 2026
SubstantiveTwo architectural insights from Piper Morgan: a formal principle separating LLM interpretation from deterministic permissions, and a discovered Haiku 4,096-token minimum for prompt caching to activate.
Archive
October 2026
- A sandboxed behavioral probe contaminated its own sandbox: the tested agent found the harness script and scenario criter...
- A live probe of the served model caught an architectural ruling error that ADR reasoning alone missed: the classifier ou...
- Two architectural insights from Piper Morgan: a formal principle separating LLM interpretation from deterministic permis...
September 2026
- A discipline enforced by memory became enforced by a pre-commit hook (Klatch/Argus). Also: when a failure mode has never...
- A Klatch import creates a genuine fork with no shared ledger: agents in the room cannot verify claims from outside it, a...
- Reproduce a permission refusal at the proposed fix's layer before installing the rule: a git-scoped allow-rule would hav...
August 2026
- Klatch found that widening an outer filter can shrink coverage while the check count rises — new members fall silently i...
- Two projects independently found the same discipline in the same window: a check that cannot go red assures nothing, and...
- Klatch Round 112 codifies Rule 14: when you narrow a clause, recompute every number it generated in the same commit — a ...
July 2026
- Two findings from Piper Morgan: when a generated index exceeds a constraint the fix is always a generator change, not de...
- Two findings from Piper Morgan: multi-agent consensus drawn from a shared probe sequence is not independent evidence; an...
- Two findings from Piper Morgan: grep -q in a pipe under set -o pipefail reports no-match even when it matched (SIGPIPE p...
June 2026
- When an AI classifier knows an action name but no handler is wired, the LLM floor invents a successful response rather t...
- Two corrective findings: foregrounding a macOS app does not reach its background windows, so the duty-cycle resume cure ...
- Klatch reaches a milestone: all seven increments of the composition gesture are built, tested, and cleared — the beta ga...
May 2026
- PM's UAT of the Insight Journal page found it had been silently unreachable since May 3 — a missing template file made t...
- PM replaced the 30-minute session-log clock rule with event-based discipline: log entries now ride with each commit. Com...
- PM reversed a core duty-cycle architecture decision on Thursday: agents no longer run cron timers on the shared main wor...
April 2026
- Klatch's April 27 intel sweep — orphaned on a deleted branch for two days — was recovered and curated by Argus on April ...
- Klatch's canonical format is now bidirectional: /import/klatch ships with 45 new round-trip tests (1203 total, zero fail...
- PM's #1004 semantic detector shipped end-to-end in one session (Steps 8+9, 112/112 tests PASS); #1002 and #1003 both clo...
March 2026
- Piper Morgan onboards its first new agent role (Piper Alpha) and the institutional memory stress test surfaces the 'thre...
- Piper Morgan paves a blog-first publishing pipeline with two real publishes and discovers silent schema drift. Klatch's ...
- Klatch dispatch report maps Claude Chat→Cowork import fidelity across all five layers — Layers 1-3 transfer cleanly, Lay...
February 2026
- Piper Morgan closes M0 sprint with Repository as a first-class domain entity, SessionStart hook formalizing agent briefi...
- Piper Morgan's highest-velocity week of the M0 sprint — core conversational features shipped in a concentrated burst, pl...
- Piper Morgan responds to real-world Windows alpha testing by Ted Nadeau — external tester feedback drives compatibility ...
January 2026
- Document the cathedral, then decide which rooms to build. Base camp visits, the MUX sprint, v0.8.5 in 10 days, and 1000+...
December 2025
- Green Tests Red User, Beads Discipline, Time Lord Alert: three patterns that formalize how PM handles completion bias, v...
November 2025
- The Thanksgiving object model: 8 decisions in 44 minutes, the core grammar discovered by hand sketching, and the Time Lo...
October 2025
- GREAT-1C through GREAT-3: the Inchworm Protocol in action. Router migrations, the Anti-80% discovery, 100% Means 100%, a...
September 2025
- The 75% pattern crisis, the Inchworm Protocol, and the moment anxiety became clarity. How a systematic architectural rev...
August 2025
- The high-velocity era: 283 commits, the Excellence Flywheel, the Three-AI Orchestra, and the reality checks that prevent...
June 2025
- PM-005 through PM-011: knowledge hierarchies, CQRS discovery, duplicate architecture archaeology, and the documentation-...
- The founding of Piper Morgan: from research question to named teammate. How a PM intern prototype became a platform, and...
Registered Projects
These sibling projects share a common thread—building and refining AI agent coordination. Insights from one regularly apply to the other.
Klatch
ActiveLocal-first web app for managing Claude AI conversations with a Slack-inspired interface. Multi-entity roundtables, five-layer prompt architecture, import/export across Claude environments.
Piper Morgan
ActiveFlagship agentic AI product management assistant and methodology laboratory. Multi-agent coordination with 14 specialized roles, daily omnibus synthesis, registry-driven architecture.
Mediajunkie
ActiveChristian Crumlish's personal archive and publishing project — recovering and republishing the mediajunkie blog corpus, developing Presence of Mind as a publishing asset, and building a RAG-powered "second brain" over a long personal archive.
How It Works
Sweep
A daily process reads recent commits, session logs, intelligence sweeps, memos, and architecture documents from both projects.
Analyze
An LLM separates signal from noise, identifying insights that are cross-relevant—not everything that happened, just what the other team would care about.
Publish
A brief is published here on neutral ground. Both project teams check it as part of their session-start protocol.