dream

A command that combines the latest links in a continuum chain into one summary link while moving the original links to an archive.

In plain words
What is it for?
Use it to roll up recent project decisions, notes, or events into a durable phase summary while preserving the original links for reference.
Why use it?
It keeps a growing record smaller without deleting its history or losing how decisions changed over time.

Command

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add commands/devzonayed/mochi/dream
Clone the repo
git clone --depth 1 https://github.com/DevZonayed/Mochi
Per session 25 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 897 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00025 $0.00897
Opus 5 $0.00013 $0.00449
Sonnet 5 $0.00005 $0.00179
Haiku 4.5 $0.00003 $0.00090

Measured yesterday against content hash 5a3702e3f746, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

dream scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

plugins/continuum/commands/dream.md · 75 lines

How it starts

The opening of the file, as written. The whole thing — 75 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are performing a rollup / dream on the continuum chain: collapsing the last N links into a single durable "phase digest" so the chain stays sub-linear in size while preserving decision lineage.

What to do

  1. Gather the input package with the helper:

    node "$(cat .continuum/.plugin-root)/lib/dream_prepare.js" --n ${ARGUMENTS:-25}
    

    Read the JSON output. Key fields:

    • rollupCandidateIds — ids that will be collapsed
    • summaries — id → summary text (this is your raw material)
    • tagsUnion — superset of tags across candidates (use to choose digest tags)
    • tsRange — first/last timestamp covered
    • canProceed — if false, ABORT and tell the user why (from warnings)
  2. Compose the digest summary in your head (~800 tokens). Rules:

    • Merge duplicates. If link 0007 says "use MFA" and link 0009 says "use MFA via TOTP," collapse to one decision with the concrete value.
    • Drop superseded-and-irrelevant. If link 0005 said "basic login" and link 0010 superseded with "MFA," the digest records ONLY MFA — but explicitly notes the supersession lineage ("MFA, supersedes earlier basic-login decision from this phase").
    • Preserve decision lineage. Never silently drop a contradiction; state it as resolved.
    • References, not payloads. If detail belongs in code/commits/archives, reference it — don't restate.

    Use the structure:

    ## Phase digest: <ids> (<tsRange>)
    
    ## Resolved decisions
    - <decision> <lineage if superseding>
    
    ## Carried-forward open threads
    - <still-unresolved at end of phase>
    
    ## Constraints / Do not (cumulative)
    - <durable "do not" items>
    
    ## Dropped (no longer relevant)
    - <briefly: superseded items not worth carrying>
    
    ## Provenance
    - rolled-up links: <ids>
    
  3. Commit the rollup atomically with the finalize helper:

    cat <<'CONTINUUM_DIGEST_EOF' | node "$(cat .continuum/.plugin-root)/lib/dream_finalize.js" --rollup-ids "<comma-separated-ids>" --tags "phase-digest,<other-tags-from-tagsUnion>"
    <YOUR DIGEST MARKDOWN HERE>
    CONTINUUM_DIGEST_EOF
    

    The script prints the new digest link id. It:

    • Writes the new digest link
    • Moves all rollup-ids directories from chain/links/NNNN/ to chain/links/_archived/NNNN/
    • Appends a tombstone entry per archived link to index.jsonl (originals are NEVER mutated)
    • Clears any pending-checkpoint sentinel
  4. Do NOT regenerate STATE.md as part of dream. STATE.md is regenerated by /continuum:checkpoint, not rollup — rollup is purely a chain-compression operation that preserves history.

  5. Report back to the user: new digest id, range covered, count rolled up, and whether you noticed any contradiction worth surfacing.

Read the full file on GitHub · 75 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 75 lines · 25 tokens per session scan A 5a3702e3f746

Subscribe to this mod's changes

dream is a command published in the GitHub repository DevZonayed/Mochi (3 stars, last pushed 15d ago), licensed MIT. It adds 25 tokens to every session and 897 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.