A test tool for checking dreamd’s memory system. It runs automated tests for saving, recalling, ranking, and identifying memories across coding agents.

Use it to run dreamd’s plumbing and quality test suites, including checks for memory ranking, source attribution, and promotion.

It shows whether dreamd’s memory pipeline works and whether important memories are returned in useful order.

How botzrDev/dreamd configures Claude Code

Reuse

Borrowing it

Nothing to install: this file belongs to botzrDev/dreamd. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file

curl -O https://raw.githubusercontent.com/botzrDev/dreamd/main/.claude/skills/alpha-test/SKILL.md

Clone the repo

git clone --depth 1 https://github.com/botzrDev/dreamd

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/skills/botzrdev/dreamd/alpha-test/github.svg)](https://agentmods.dev/skills/botzrdev/dreamd/alpha-test)

Your own site

<a href="https://agentmods.dev/skills/botzrdev/dreamd/alpha-test"><img src="https://agentmods.dev/badge/skills/botzrdev/dreamd/alpha-test/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/skills/botzrdev/dreamd/alpha-test"><img src="https://agentmods.dev/badge/skills/botzrdev/dreamd/alpha-test.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 108 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 2,094 The whole file, excluding the scripts and references it only reads on demand.

Security scan C 1 finding. 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	$0.00108	$0.02094
Opus 5.5	$0.00043	$0.00838
Sonnet 5.5	$0.00022	$0.00419
Haiku 4.5	$0.00011	$0.00209

Measured 1mo ago against content hash f03ceda87463, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade C, and why

alpha-test scanned grade C with 1 finding 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.

Recursive force deletehighDestructive command

rm -rf with a variable or a broad path is one typo away from removing the wrong tree.

`rm -rf "$SANDBOX"` in the `cleanup()` trap, re-run, then read

.claude/skills/alpha-test/SKILL.md · 125 lines

How it starts

The opening of the file, as written. The whole thing — 125 lines — stays where its author put it; the contents beside it link to each section on GitHub.

dreamd alpha test

Two automated suites live at scripts/alpha/:

Plumbing (alpha-suite.sh) — proves a learning appended by one harness is recalled by an independent second harness, on both the daemon path (Phase 2) and the no-daemon JSONL-replay path (Phase 1). Uses lexically-unique payloads, so it proves recall happens but says nothing about whether it's good.
Quality (quality-suite.sh + optional quality_judge.py) — proves the memory is actually good through the MCP boundary: salience-weighted ranking, attribution, and dream-cycle promotion (deterministic gate), plus a fuzzy natural-language relevance report (LLM judge). See Quality suite below.

Both are the code-path proof only; the real GUI round-trip (actual Cursor / Claude Code MCP clients + screenshot) is the manual DEMO-4 runbook, out of scope.

Run

cargo build -p dreamd            # both suites run target/debug/dreamd
scripts/alpha/alpha-suite.sh     # plumbing — pass = "7 passed, 0 failed"
scripts/alpha/quality-suite.sh   # quality golden gate — pass = "14 passed, 0 failed"
scripts/alpha/quality_judge.py   # optional LLM-judge relevance report (needs an API key)

Pass = exit 0 and 7 passed, 0 failed. Needs python3 and git on PATH. The suite redirects HOME=$(mktemp -d), so your real ~/.agent daemon, registry, and memory are never touched — it kills the daemon and deletes the sandbox on exit.

The 7 checks (in order)

Phase 2 (daemon up): (1) daemon binds its socket; (2) a claude-code append mints an evt_… id; (3) a separate cursor process recalls that write (polls ~18s for the daemon's ~5s index-commit cadence).
Boundary: (4) socket is gone after the daemon is killed.
Phase 1 (no daemon): (5) a cursor append mints an evt_… id; (6) a fresh process replays the JSONL and recalls it; (7) that replay also recalls the earlier Phase-2 write.

Diagnosing a failure

The sandbox is deleted on exit. To inspect logs, temporarily comment out rm -rf "$SANDBOX" in the cleanup() trap, re-run, then read $SANDBOX/daemon.log (the script prints the sandbox path on the first line).

Read the full file on GitHub · 125 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 First seen · 125 lines · 108 tokens per session scan C f03ceda87463

Subscribe to this mod's changes

alpha-test is a skill published in the GitHub repository botzrDev/dreamd (6 stars, last pushed yesterday), licensed Apache-2.0. It adds 108 tokens to every session and 2,094 once invoked, about $0.0004 per session on Opus 5.5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

LLM Memory and context Testing ai-agents coding-agents local-first mcp

remember

A memory-saving skill that stores facts, preferences, decisions, tasks, and other context in Super Memory, a local semantic memory store.

Utundry/sloplesscode · 60 tokens

recall

A memory-search skill for finding previously saved information in Super Memory, a semantic memory store.

Utundry/sloplesscode · 52 tokens

dev-browser

Browser automation with persistent page state. Use when users ask to navigate websites, fill forms, take screenshots, extract web data, test web apps, or automate browser workflows. Trigger phrases include "go to [url]", "click on", "fill out the form", "take a screenshot", "scrape", "automate", "test the website"…

MemTensor/MemOS · 84 tokens

knowledge-base

Skill "knowledge-base" from Pinvou/pinvou-agent, covering ima knowledge base, operations, routing rules, examples and file upload guard.

Pinvou/pinvou-agent · 0 tokens

notes

Skill "notes" from Pinvou/pinvou-agent, covering ima notes, operations, write rules, examples and response handling.