A command for reviewing proposed changes stored in a dream artifact, a saved set of planning or memory files, and applying only the changes accepted one by one. It first validates the artifact and its files.

Use it to validate a dream artifact, inspect its proposals and apply selected roadmap or memory changes. It accepts a timestamp, or can resolve the newest artifact with `latest`.

It prevents malformed, unsafe or unreviewed proposals from being applied automatically. It also checks that the memory repository is clean before the review begins.

dream-apply — Command for Claude Code by conorbronsdon

Install

Getting it into your agent

It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.

Clone the repo

git clone --depth 1 https://github.com/conorbronsdon/agent-context-os

agentmods

npx agentmods add commands/conorbronsdon/agent-context-os/dream-apply

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/commands/conorbronsdon/agent-context-os/dream-apply/github.svg)](https://agentmods.dev/commands/conorbronsdon/agent-context-os/dream-apply)

Your own site

<a href="https://agentmods.dev/commands/conorbronsdon/agent-context-os/dream-apply"><img src="https://agentmods.dev/badge/commands/conorbronsdon/agent-context-os/dream-apply/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/commands/conorbronsdon/agent-context-os/dream-apply"><img src="https://agentmods.dev/badge/commands/conorbronsdon/agent-context-os/dream-apply.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 20 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.

When invoked 3,273 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

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.

Model	Per session	Once invoked
Fable 5.1	$0.00020	$0.03273
Opus 5.5	$0.00008	$0.01309
Sonnet 5.5	$0.00004	$0.00655
Haiku 4.5	$0.00002	$0.00327

Measured 1mo ago against content hash 4d92d3b919ec, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

dream-apply 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 1mo ago.

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.

.claude/commands/dream-apply.md · 188 lines

How it starts

The opening of the file, as written. The whole thing — 188 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/dream-apply — review + apply a curator pass

Substrate background: docs/dream-architecture.md.

Usage

/dream-apply {ISO-timestamp}
/dream-apply latest        # auto-resolves to most recent .dreams/ subdir

Steps

1. Resolve the dream dir

Run the executable validator before reading or applying any artifact:

python3 scripts/dream/validate-memory.py artifact "${ARGUMENTS:-latest}"

Parse the returned JSON. Set MEMORY_DIR from memory_dir, TS from timestamp, and the dream dir from dream_dir. The helper rejects malformed timestamps, path traversal, absolute or control-character arguments, symlinked artifact components, missing proposals.json/REPORT.md, malformed or colliding proposal schemas, unknown/roadmap actions, empty evidence, unsafe control-file targets, and any proposal filename outside the validated memory root. It also requires the memory repository to be clean before review; stop so any host auto-memory change can be reviewed or snapshotted separately.

Do not create or guess a memory directory when a check fails. Stop and direct the user to docs/auto-memory.md.

If $ARGUMENTS is latest or empty, the helper resolves the most recent valid dream artifact by name. Do not reimplement timestamp/path selection in prose.

2. Load the proposal artifact

Read the validated $dream_dir/proposals.json and $dream_dir/REPORT.md. If either becomes missing or changes after validation, rerun the helper and stop on any error.

3. Show the report header

Print the top of REPORT.md (header + counts). Don't dump the whole thing — the user already has it if they want the full text.

4. Walk each proposal

For each proposals[i]:

a. Print a header, then the fields relevant to the action:

─── Proposal {id} ({i+1}/{N}) ───
Action: {action}    Confidence: {confidence}
Reasoning: {reasoning}
Evidence:
  - {evidence[0]}
  - {evidence[1]}

modify (content): Target: {target} then Current: / Proposed: excerpts.
archive: Target: {target} then the one-line archive reason.
merge (structural): Absorb: {targets} → Survivor: {survivor} (net index lines: {net_index_lines}), then the merged_body and the index_changes.
split (structural): Target: {target} → Into: {result_files[*].name}, then each child's purpose + index_line and its body.
add / flag: target + proposed content (add) or the flagged concern (flag).

Read the full file on GitHub · 188 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.

1mo ago Changed 4d92d3b919ec
1mo ago First seen · 188 lines · 20 tokens per session scan A cb10230540d9

Subscribe to this mod's changes

dream-apply is a command published in the GitHub repository conorbronsdon/agent-context-os (30 stars, last pushed yesterday), licensed MIT. It adds 20 tokens to every session and 3,273 once invoked, about $0.0001 per session on Opus 5.5. 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 2026-08-30.

Claude Code Planning ai ai-context context-management developer-tools template

orchestrate

Do this task as an orchestrator. The aim is the same result at a fraction of the cost: in a long session, every token in this conversation is re-read on each later request, while a subagent's reading is discarded when it returns. The plugin's subagents run on Haiku 5.5, about 40× cheaper per token than Opus 5.5.

egorfedorov/claude-context-optimizer · 0 tokens

dream-apply

Getting it into your agent

Wrote this? Show the measurements

What it costs to keep this loaded

Grade A, and why

Nothing flagged

How it starts

/dream-apply — review + apply a curator pass

Usage

Steps

1. Resolve the dream dir

2. Load the proposal artifact

3. Show the report header

4. Walk each proposal

What this file has done since we first saw it

Other commands, from other repositories

orchestrate

generate-docs

workflow:ui-design:reference-page-generator

sync

handoff

review-code