agent-knowledge

The readable memory of T's daily-dev-agentic loop: what the agent learned from daily.dev, distilled into lessons that name why each one matters for the stack (Claude Code, MCP, multi-agent systems, LangGraph, Firebase, Python, Rust, local LLMs, fintech infrastructure) and what to do about it.

Written mostly by a machine. The daily.dev MCP connector (Cloudflare Worker dailydevplugd) appends lessons after each scheduled run; Claude sessions add verified fact sheets; humans fix typos. Run state (seen-post watermark, counters) lives in Cloudflare KV, not here.

Status (2026-09-27): the lesson feed is paused. The scheduled run ("daily-dev-agentic · daily learn") never had the daily.dev connector attached, so from 2026-08-24 it exited early every day while reporting success; T retired it on 2026-09-27. The last lessons are from 2026-08-24. To restart: add the connector in claude.ai (Settings → Connectors → custom connector), enable it on that scheduled task, and switch the task back on.

Layout

Path	What it holds	Written by
knowledge.md	Every lesson, newest day first. Each links its daily.dev source post and carries a confidence score	Connector (knowledge_append)
lessons/YYYY-MM-DD.md	The same lessons, one file per day	Connector
verified/YYYY-MM-DD_topic_vN.md	Fact sheets verified live in a Claude session, not daily.dev lessons. Revisions bump vN	Claude sessions, hand-maintained
harness/check-knowledge.py	PASS/FAIL check that the files above still have the shape readers rely on	Run by CI and by sessions
harness/knowledge_md.py	The one Markdown parser the checks use: markdown-it-py (CommonMark 0.31.2) plus the patches that align it with the reference	Sessions
harness/tests/	316 negative-test cases: each copies the repo, breaks one thing, and asserts the checker's verdict	Sessions
harness/conformance/	Differential gates: the parser must render all 652 CommonMark spec examples and 71 edge cases like the reference implementation, and the HTML tokenizer must split HTML like html5lib	Sessions
LEDGER.md	One row per hand-made or session-made change (connector runs are not logged here)	Sessions and hand edits

How a lesson is shaped

## 2026-08-24

- **One actionable sentence, supported by the article.**
  - Why it matters here: the link to T's stack or a live project.
  - Do: the concrete next action, or "None now".
  - Source: [Title](https://...) — Author (⬆️ upvotes · 💬 comments) · tags: a, b · confidence 80%

Confidence follows the connector's scale: official docs or maintainer posts around 80–95, well-upvoted practitioner posts 60–80, single opinions 30–55. Zero lessons on a quiet day is a valid result; the loop is told never to pad.

Reading it

• Newest first: open knowledge.md and read from the top.
• One day: open lessons/<date>.md.
• Something proven rather than read: verified/.
• Programmatically: the connector's knowledge_read tool (days: N) returns the same content without cloning.

Checking it

pip install -r harness/requirements.txt     # markdown-it-py, pinned
python3 harness/check-knowledge.py          # PASS/FAIL lines
python3 harness/check-knowledge.py --json   # same, as one JSON object
bash harness/tests/run.sh                   # the 316 negative-test cases

The checker reads Markdown the way CommonMark 0.31.2 does, using a real parser rather than pattern matching. Two copies of a lesson count as the same when they render to the same visible HTML. Lessons are plain Markdown: raw HTML can hide, restyle or swallow what a reader sees, so a lesson holding any raw HTML other than a <!-- comment --> fails the check. CI runs the checker, the test suites, markdownlint, shellcheck, pyflakes and two conformance gates (the Markdown parser against the reference implementation, the HTML tokenizer against html5lib) on every pull request and every push to main (.github/workflows/lint.yml). GitHub itself renders with an older spec (cmark-gfm, 0.29): in rare edge cases, such as a lowercase <!foo> line, GitHub shows text where 0.31.2 hides a declaration.

Hand edits

Safe. The connector only inserts and de-duplicates by source link; it never rewrites what is already there. Conventions for edits, fact sheets and the ledger are in CONTRIBUTING.md.

License

MIT. See LICENSE.