agentleFS
NVIDIAKnown publisherNVIDIA/nemoclaw-community/examples/recipes/nvidia/memory-driven-chief-of-staff/profile/skills/memory-consolidation/SKILL.md

memory-consolidation

Skill142 starsChanged 3 months agoApache-2.0View on GitHub

Keep the memory inside its size limits by compacting rather than truncating, so it stays readable as it ages.

5 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

9 rules. 0 are common, 4 are distinctive, and 5 are very specific to this project.

  • 0

    Common

    Most of the 5 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    All reads of workspace/memory use python3 $HERMES_HOME/scripts/memory_operations.py read <relative-path> ....

    line 15
  2. Distinctive:

    If it reports a pending or blocked operation, stop this pass and surface the diagnostic.

    line 18
  3. Very specific:

    All memory file changes, including shared index entries and the single log entry, go through python3 $HERMES_HOME/scripts/apply_memory.py < …

    line 21
  4. Very specific:

    Never write, append, rename, or delete memory files with shell or file tools.

    line 23
  5. Very specific:

    Use null text for a reviewed people merge source deletion and include its index removal in the same operation.

    line 26
  6. Very specific:

    Before building a proposal, read $HERMES_HOME/scripts/memory-protocol.md for the installed JSON contract and examples.

    line 36
  7. Distinctive:

    Read the schema's growth-control table first.

    line 52
  8. Distinctive:

    Provenance on every claim that survives: If you keep the claim, keep its footnote.

    line 74
  9. Distinctive:

    If a page needs judgment you cannot make safely — two entries that might be the same project, a commitment you cannot tell is resolved — …

    line 86

“Common” is measured: kept by at least 6 in 10 of the 5 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Coordinated memory transport

All reads of workspace/memory use python3 $HERMES_HOME/scripts/memory_operations.py read <relative-path> .... This returns the store instance, exact file hashes, and a coherent snapshot. If it reports a pending or blocked operation, stop this pass and surface the diagnostic. Do not read partial files as current memory.

All memory file changes, including shared index entries and the single log entry, go through python3 $HERMES_HOME/scripts/apply_memory.py < proposal.json. Never write, append, rename, or delete memory files with shell or file tools. Prepare complete people/attention pages and their evidence markers together in a legacy_write proposal with the returned expected_hash for every file. Use null text for a reviewed people merge source deletion and include its index removal in the same operation. Preserve the content and admission rules below. A failed proposal acknowledges no new evidence batch.

Managed projects, patterns, and concepts use update with declared field changes; never send a managed page through the whole-file adapter. Existing unmanaged pages may be repaired with their exact full-file precondition and remain unmanaged. Use repair_index for verified missing or stale entries, and log_only for a pass that changes only the shared memory log. The writer constructs index patches and a log entry containing its operation ID. Before building a proposal, read $HERMES_HOME/scripts/memory-protocol.md for the installed JSON contract and examples.

A conflict preserves the live file. Report it and defer that target. Generate a fresh independent proposal for unrelated work; do not retry a changed request under an already used request ID, adopt a page implicitly, clear review flags, or use a direct write as a fallback. Per-page user approval is not required for generated content after Foundation enablement. Handwritten content requires an explicit scoped user action through correct.py.

Consolidating the memory

Memory that only grows stops being read, and a memory nobody reads is worse than none — it looks like the system knows things it can no longer surface. This job keeps pages inside the ceilings in $HERMES_HOME/schema.md.

Read the schema's growth-control table first. Work only on pages the repair job reported as over their ceiling, or that you can see are over.

Compact, do not truncate

The distinction is the whole job.

  • Dated events move, they do not vanish. A project's dated bullets belong in that project's log.md; the main page keeps the current picture.
  • Superseded state is dropped. "Waiting on the vendor to reply" is dead once the vendor replied. That is not information loss, it is the page catching up.
  • A person's Recent Interactions past 30 items become a Relationship Arc: a few sentences on how the working relationship changed. Write it once, revise it later, never append to it.
  • Patterns decay. A behaviour last observed months ago and never since is weakened or dropped, not restated with confidence it no longer has.

What must survive

  • Unresolved commitments. Anything the user owes someone, or is owed, survives compaction regardless of age. Age is not resolution.
  • Provenance on every claim that survives. If you keep the claim, keep its footnote. A compacted page full of unsourced assertions is worse than the long one it replaced.
  • Project history. Moved to log.md, never deleted. Rotate the log at its limit rather than trimming it.

Bounded work

Compact the pages that are over, and stop. Rewriting the whole memory because two pages were long is how a maintenance job turns into an outage — and every rewrite is an opportunity to lose a footnote.

If a page needs judgment you cannot make safely — two entries that might be the same project, a commitment you cannot tell is resolved — leave it and say so in the log. An honest deferral is a good outcome.

Log

## [2026-08-18T09:31:00Z] consolidation
- projects/billing_migration: moved 14 dated bullets to log.md, page now 4 sections
- people/dana_okoro: Recent Interactions 41 -> 30, wrote Relationship Arc
- deferred: two entries under concepts/ may be the same term, evidence unclear

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.