The briefing
Phantom pull requests recovered offline posterior update, replay, regret and decision reader functions from an open PR. Other changes added lane health badges, spend down data, reply outbox and new ask types.
Phantom PRs recovered pure functions for offline posterior update, IPS and doubly robust replay, regret computation, decision reader and posterior comparison. The pull requests state these are bounded recoveries from PR 5952 which remains open.
Phantom added lane health verdicts and badges. The description says laneHealthFor returns states including ok, held, starved, wasteful and drained with precedence rules and a wasteful threshold of five starts per merged PR.
Phantom added hourly leader replies to the outbox via a new phm outbox from-log command. The pull request says it appends deduplicated replies to MASON-OUTBOX.md.
Phantom added pools.spenddown.v1 derived from pools.v1. The description states it supplies even-pace and projected fractions plus a chart tick for phone and desktop surfaces.
Phantom extended asks with usage and approval kinds. The pull request says tagged lines now carry subject, question and unblock text.
Ashlr Stack PRs added shared JSON error output for stack list and status commands plus README schema docs. Another updated test guidance and CLI exit code coverage.
Ashlr MD and Phantom Secrets PRs updated contribution guidance and changelog checks. The summaries state validation runs daily and local CI skips some checks.
Review the recovered offline functions in coding-cli and scheduler modules. Check lane-badge.ts and pools-spenddown.ts for new data shapes.
Lane health and spend down data give clearer visibility into fleet state and usage pace.
Pull latest Phantom and Ashlr Stack changes. Run the documented pre-push checks listed in the updated MD and Stack guides.
Ashlr tools connected to this story. Review their source and setup requirements.
Work with eligible coding-agent accounts and local models in one workbench; review fleet proposals under signed authority.
Source material behind this briefing. Our practical implications are Ashlr’s interpretation.
This briefing was prepared with AI assistance from the linked sources. The analysis may include interpretation; consult the original sources for complete context. Cover artwork is illustrative.


Ashlr release
Phantom published releases 3.30.67 and 3.30.66. Lexicon and Locus updated documentation on test selection and adapter guidance.
EngineeringThe briefing
AI-generated illustration
The AI evaluation nonprofit added tasks that require research taste and judgment to its time-horizon evaluation. Frontier models still lag humans on these tasks even while they continue to double the length of verifiable tasks they can complete every four months.


Ashlr release
Keep coding agents scoped to the right client, account, and tenant. Manage identity bindings across command-line tools and MCP connections.
Keep a personal vocabulary of names and speech-to-text corrections in one YAML file, then use it across coding-agent tools, MCP connections, and browser workflows.
Read, edit, and export Markdown in a desktop app, with AI assistance and tools for reviewing agent-written documents. Available for macOS, Windows, and Linux.
Delegate supported API work through scoped placeholders and a local proxy, keeping real keys out of agent context.
Token-efficient tools and workflows for Codex and Claude Code. Read, search, edit, and inspect repositories with compact outputs and visible context usage.
Several Phantom pull requests updated the unified inbox to show legacy bus messages and inbox triage counts. Other changes added fleet yield, landed, and pace tools plus documentation updates across projects.