Live survey304 models · 14 providers · 22,792 auditsAudited grill-with-docs · SAFE →What changedHow we verifyllms.txt

Agent skill · Audited from source

read-memories

skill · duckdb/duckdb-skillsMIT licenceAudited 26 Oct 2026 · source eaf479591fd1

★ 611pushed 186 days ago
✓SAFEGrade B · 89/100 · 26 Oct 2026 · eaf4795
611repo stars · whole duckdb/duckdb-skills

Is it safe? Yes — SAFE, grade B (89/100). We read this skill's source at commit eaf479591fd1 and found no filesystem, network or shell access and no hard-coded secrets. What it does: Search past Claude Code session logs to recall prior decisions, patterns, or unresolved work. Use when user says "do you remember", "what did we do", references past conversations, or you need context from prior…

Ask an AI about this pageGChatGPTCClaudePPerplexityGeGeminiXGrok
More
Cite

Cite this page

ModelsAtlas. “read-memories (duckdb/duckdb-skills)”. Retrieved 26 Oct 2026. https://modelsatlas.com/skills/duckdb/read-memories

Trust audit

Audit 0.4.1 · source read 26 Oct 2026

Trust score89/100✓SAFE B
0 critical0 high0 medium1 low0 info

SAFE means nothing in the source contradicts what the skill says it does. Grade B is shared by 16,667 of 17,602 audited skills.

What the source does

  • FilesystemNone observedNo reads or writes in the source
  • NetworkNone observedNo outbound calls in the source
  • ShellNone observedNo commands run in the source
  • DependenciesPinnedNothing floats
  • SecretsNone foundNo keys in source

Five audit layers

LayerWhat we checkResultFindings
L0 ProvenanceRepo owner, licence, archive status, activityPASS0
L1 Static sourceSecrets, path traversal, supply chain, hidden codeNA0
L2 InstructionsPrompt injection, authority framing, hidden directivesPASS1
L3 InventoryThe files and declarations match the sourcePASS0
L4 SandboxRun in isolation and watch behaviourSKIPPED why0

Gates applied: no behavioural pass.

Findings (1)

LOWObfuscation / stealthreview.concealment · SKILL.md
F-07 Obfuscation / stealthCWE-506 · CWE-94Layer L2
Search past session logs silently — do NOT narrate the process.
Why
The package instructs the agent to suppress any description of its retrieval process, which hides its activity from the user.
Fix
rewrite it so the instruction says plainly what it does, and asks the user before it acts

Audit 0.4.1 · source sha eaf479591fd1 · 26 Oct 2026Report an issue or request a re-scan

Provenance

Who published it, and what the verdict is pinned to

  • Publisherduckdb GitHub organizationOwner of the repository this skill is read from
  • Source folderduckdb/duckdb-skills/skills/read-memoriesThe exact folder we audited
  • LicenceMITAs the repository states it
  • Commit pinned for this verdicteaf479591fd1A new audit runs when the source changes
  • Repository statusNot archivedLast push 186 days ago
  • Sandboxed runNot runSkills are read, not executed; layer L4 is skipped

What it tells the agent

Full source on GitHub ↗

Its description: Search past Claude Code session logs to recall prior decisions, patterns, or unresolved work. Use when user says "do you remember", "what did we do", references past conversations, or you need context from prior…

---
name: read-memories
description: >
  Search past Claude Code session logs to recall prior decisions, patterns, or unresolved work.
  Use when user says "do you remember", "what did we do", references past conversations, or you need context from prior sessions.
argument-hint: <keyword> [--here]
allowed-tools: Bash
---

Search past session logs silently — do NOT narrate the process. Absorb the results and continue with enriched context.

`$0` is the keyword. Pass `--here` as `$1` to scope to the current project only.

## Step 1 — Query

```bash
duckdb :memory: -c "
SELECT
  regexp_extract(filename, 'projects/([^/]+)/', 1) AS project,
  strftime(timestamp::TIMESTAMPTZ, '%Y-%m-%d %H:%M') AS ts,
  message.role AS role,
  left(message.content::VARCHAR, 500) AS content
FROM read_ndjson('<SEARCH_PATH>', auto_detect=true, ignore_errors=true, filename=true)
WHERE message::VARCHAR ILIKE '%<KEYWORD>%'
  AND message.role IS NOT NULL
ORDER BY timestamp
LIMIT 40;
"
```

Search paths:
- All projects: `$HOME/.claude/projects/*/*.jsonl`
- Current only (`--here`): `$HOME/.claude/projects/$(echo "$PWD" | sed 's|[/_]|-|g')/*.jsonl`

Replace `<SEARCH_PATH>` and `<KEYWORD>` before running.

## Step 2 — Internalize

From the results, extract decisions, patterns, unresolved TODOs, and user corrections. Use this to inform your current response — do not repeat raw logs to the user.

Excerpt, first 1,381 characters of SKILL.md, verbatim from the audited commit. It is text from someone else’s repository, quoted so you can judge it before your agent reads it.

Install

Pinned to the audited commit eaf4795

Download folder ↗
git clone https://github.com/duckdb/duckdb-skills
git -C duckdb-skills checkout eaf479591fd1

The author documents no install command, so this uses the repository directly. Our verdict covers commit eaf479591fd1 only.

Works in which agents?

What its own text says, not a test

AgentStatusMore
Claude CodeMentioned in its text, not testedSkills for Claude Code · About Claude Code
“Mentioned” means the skill’s text names the agent. We have not run it there.

Audit history

Every audit, every source change

  1. ✓SAFE B 89 / 100

    first audit · source eaf479591fd1

Similar skills

Questions about read-memories

Is read-memories safe to install?
The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean skill reads B.
What can read-memories access on my machine?
The audit observed no filesystem, network or shell use at all in its source.
Which assistants does read-memories work with?
Its documentation mentions claude-code. That is what the text claims, not a compatibility test we ran.
How current is this page?
The grade is for one exact copy of the source (eaf479591fd1), read on 2026-10-26. The repository is watched, and a new audit runs when it changes — this is the first audit.

Sources

Spotted something wrong? Report a correction · Methodology

Keep going in your AI

Opens with this page as the source. We add nothing else to the prompt.