docs(ops): refresh roadmap freshness + anchor on human-signal baseline#577
docs(ops): refresh roadmap freshness + anchor on human-signal baseline#577bmdhodl wants to merge 1 commit into
Conversation
Bump Last reviewed to 2026-06-05 (was 5+ days stale per CLAUDE.md threshold). Re-anchor NOW/NEXT/LATER on the human-signal-vs-machine-volume split from sdk_pulse.py (2026-06-05): PyPI published-package downloads are real signal (CI installs from local checkout, never pip install agentguard47); clone counts are partly self-inflicted by scheduled checkout workflows and are discounted. MCP distribution-hygiene success signal now measures human-signal proxies (PyPI downloads, stars, referrers), not raw clone/download volume. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
🤖 Claude reviewLGTM - no blocking issues. |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: d0ffccf496
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| they directly strengthen coding-agent adoption. | ||
|
|
||
| **Last reviewed:** 2026-05-29 | ||
| **Last reviewed:** 2026-06-05 |
There was a problem hiding this comment.
Bumping the roadmap's Last reviewed date here marks this file as fresh even though the unchanged Current Focus Notes immediately below still conflict with the tracked memory files: memory/state.md says PyPI latest is 1.2.13, the MCP Registry serves 0.2.2, and Glama has all seven tools indexed, while this roadmap still says 1.2.10, 0.2.1, and tools is empty. Because AGENTS.md requires agents to rely on the roadmap freshness check before starting work, this suppresses the stale-roadmap warning for several days and can send future work toward already-resolved blockers; either update the stale notes in the same change or don't refresh the reviewed date.
Useful? React with 👍 / 👎.
1 participant
Summary
ops/03-ROADMAP_NOW_NEXT_LATER.mdLast reviewed to 2026-06-05 (was 2026-05-29, past the 5-day freshness threshold inCLAUDE.md).sdk_pulse.py(2026-06-05): PyPI published-package downloads are real signal (CI installs the SDK from the local checkout, neverpip install agentguard47); clone counts are partly self-inflicted by scheduled checkout workflows and are discounted.Test plan
ops/03-ROADMAP_NOW_NEXT_LATER.md(+12 / -2).Artifacts
WORK_PLAN / RESEARCH / QA_REPORT produced locally by the queue-worker (not committed). QA verdict: PASS. Source: vault
Knowledge/syntheses/agentguard-distribution.md§ Self-inflicted-CI audit (2026-06-05).Risk
Low. Docs-only, single file, zero runtime impact.