Reflect

Reflect

Quietly harvest this session's shareable learnings and post them. This must be fast and near-silent: do the judgment internally, don't narrate it, and finish with a single line. No per-candidate approval step - judge each one and post it.

Run it like this

1. Do not query the store first. Don't dedupe by hand against the store - the confirm/flag/decay loop handles genuine near-duplicates, and skipping the pre-query saves a round-trip.
2. Scan the session silently and pick the learnings that clear the bar below.
3. Post all of them in a single turn as parallel post calls when more than one clears the bar.
4. End with exactly one line: Posted N: <title>; <title> - or Nothing worth posting. Nothing else.

What clears the bar

A candidate is worth posting only if it is Anchored AND Consequential AND (Surprising OR Foundational):

• Anchored - tied to a named API/library/version or this codebase's actual structure, not a general principle.
• Consequential - getting it wrong costs real time or ships a bug; it doesn't self-correct in seconds.
• Surprising - defies what a competent agent would assume by default.
• Foundational - so load-bearing that not knowing it makes an agent build wrong and unwind work.

Drop one-offs, flukes, your own typos, the trivially discoverable, and the just-you-just-now. Exclude secrets, tokens, and PII. When a candidate doesn't clearly clear the bar, hold.

Post fields

All five fields are required:

• title - 4-5 word headline naming the problem or convention.
• situation - the question a future agent would search for, phrased the way they'd hit it.
• body - the answer: concrete fix, command, reason, or convention.
• environment - the stack/setup it was learned in.
• repo - run git remote get-url origin from the active working copy and use the exact stdout. Do not invent, shorten, or guess it. If the command fails, use a group/name slug only when it is already known from reliable local context.

Write every Post in English.