aeonfun/Memory Flush

> ${var} — Topic to focus on. If empty, flushes all recent activity.

If ${var} is set, only promote entries related to that topic. Pruning (step 3) and the timestamp/index/rotation upkeep (steps 4, 7) still run globally — a focused flush must never leave the rest of the store stale.

Read memory/MEMORY.md for current memory state.

Scan window — read logs since the last consolidation, not a fixed 3 days. Look at the *Last consolidated: <date>* line near the top of MEMORY.md:

• A real date → read every memory/logs/ file dated on or after that date (re-reading the last consolidated day is fine; the dedup check in step 2 makes it idempotent).
• never, missing, or unparseable → read the last 3 days.
• A gap longer than 14 days (agent was dark) → read the last 14 days only and note the gap in your step 6 log. Older un-promoted entries are unrecoverable; don't pretend otherwise.

This closes two holes in the old fixed 3-day window: entries older than 3 days were lost whenever the agent skipped runs, and a daily schedule re-scanned the same 3 days every time.

Steps

1. Scan the in-window logs for entries worth promoting to long-term memory

• New lessons learned (errors encountered, workarounds found)
• Topics covered (articles, digests) — add to the recent output/articles/digests tables
• Features built or tools created
• Important findings from monitors (on-chain, GitHub, papers)
• Ideas captured that are still relevant
• Goals completed or progress milestones

2. Check each candidate against existing MEMORY.md content — dedup precisely

Skip if already recorded. Dedup by the fact's subject, not by string match:

• Identify what each candidate is about (a skill, a token, a repo, a lesson, a priority).
• If MEMORY.md already carries that subject, edit the existing line in place (merge the new detail, bump any date). Never append a second bullet that paraphrases an existing one — that near-duplicate drift is what a memory flush exists to prevent.
• Only add a new bullet when the subject is genuinely absent.

3. Remove stale entries — this is as important as adding new ones

a. Open Improvement PRs section: Run gh pr list --state open --search "improve:" --json number,title,url and compare against any "Open Improvement PRs" section in MEMORY.md.

• If all listed PRs are now merged/closed, remove the section entirely.
• If some PRs are merged, update the list to reflect only current open ones.

b. Next Priorities section: Cross-check each listed priority against recent logs and current repo state. Remove priorities that are already done (e.g., "Merge open PRs" if 0 open PRs exist). Add any newly urgent priorities surfaced by recent logs. c. Lessons Learned: Remove lessons that are now outdated or resolved (e.g., a workaround for a bug that was later fixed). d. Overflow any section that outgrows its budget (keeps MEMORY.md an index, not a ledger): if a section has grown past the last ~10–15 rows — the Skills Built table is the usual first offender, but the rule is general — archive the oldest rows to memory/topics/<section>-history.md (e.g. skills-history.md) and leave a one-line pointer to that file. Trim newest-kept, oldest-archived.

4. Update memory

• Add brief entries to MEMORY.md (keep it under ~50 lines as an index).
• If a topic needs more detail, write to memory/topics/<topic>.md instead (see step 7 — register it in the index).
• Update tables (recent articles, recent digests) with new rows.
• Before adding a section, check whether its ## Heading already exists anywhere in MEMORY.md — if it does, update that section in place. Never prepend a duplicate heading.
• Stamp the consolidation date. Update the *Last consolidated: <date>* line near the top to today's date (${today}). If the line is missing, add it directly under the title. This drives the step-scan window above and is how other skills (e.g. action-converter) tell a live, consolidated store from an untouched template — leaving it at never after a real flush is a bug.

5. Make targeted edits only

Do NOT rewrite the whole file — make targeted additions and removals.

6. Register any new topic files in the index

If step 4 created a new memory/topics/<topic>.md (or a *-history.md archive in step 3d), add a one-line pointer to it under the # Reference section of memory/topics/index.md, matching the existing row format. New topic notes that aren't linked from the index become orphans no other run can find.

7. Rotate the log journal so memory/logs/ stays bounded

memory/logs/ is append-only and grows one file per day forever. Once its durable content has been promoted, roll old dailies up so the directory doesn't grow without limit — content-preserving, never a bare delete (rm isn't granted; use git rm):

• Only act when there are more than ~45 daily files.
• Pick whole calendar months that are entirely older than the current 14-day scan window (so nothing still in scope is touched). For each such month, append its dailies in date order into memory/logs/archive/YYYY-MM.md (create the dir if needed), then git rm the dailies you just rolled up.
• The archive preserves every line — this respects the append-only contract while bounding the file count. Never touch a log inside the scan window, and never edit a daily's existing content.

8. Log to memory/logs/${today}.md

Log what you promoted, pruned, archived, and the scan window you used (start date → today). If a >14-day gap was clamped, say so.

If nothing worth promoting or removing and no rotation was due, log MEMORY_FLUSH_OK and end (still stamp the Last consolidated date in step 4 — a clean flush is still a consolidation).

Network note

gh pr list uses the gh CLI's built-in auth — no curl env-var expansion. All other work is local file I/O against memory/ (plus git rm for log rotation).

Constraints

• Keep MEMORY.md an index (~50 lines). Detail lives in memory/topics/.
• Never duplicate an existing ## Heading or an existing fact — update in place.
• Pruning stale entries is as important as adding new ones.
• This skill owns MEMORY.md consolidation. Other skills (e.g. self-improve) may flag memory-hygiene problems, but structural pruning and archiving of MEMORY.md should land here to avoid two skills thrashing the same file. If self-improve prunes in an audit, treat it as a stopgap, not a reason to skip the next flush.