Skills ai-maestro-memory-search
📦
ai-maestro-memory-search
Content revision r1 High Risk ⚙️ External commands📁 Filesystem access
Search Prior AI Conversation Memory
AI agents lose context across sessions and may repeat past work. This skill directs Claude, Codex, and Claude Code to search prior conversations before acting.
Supports: Claude Codex Code(CC)
Stop for confirmation before installing.
Review the plan and obtain explicit user consent before changing files.
Install with my Agent
Copy this request to your Agent. It includes the canonical Skill page and manifest.
Agent request
Review the Skillstore skill "ai-maestro-memory-search" from https://skillstore.io/skills/23blocks-os-ai-maestro-memory-search.md and its manifest at https://skillstore.io/api/skills/23blocks-os-ai-maestro-memory-search/manifest. Verify the artifact. Stop and obtain explicit user consent before installing or changing files.Your Agent should still show its plan and request any confirmation required by the security policy.
Agent-readable resources
Use these links when an AI agent, crawler, or script needs clean context instead of reading the full page.
Test it
Using "ai-maestro-memory-search". User asks: Continue the authentication work from last time.
Expected outcome:
Found prior decisions about session storage and remaining validation work. Start by checking the existing branch and tests.
Using "ai-maestro-memory-search". User asks: Fix the issue we discussed with payments.
Expected outcome:
- Prior context points to a retry failure after payment authorization.
- The next step is to inspect the payment service logs and related tests.
Using "ai-maestro-memory-search". User asks: Use the approach we agreed on for error handling.
Expected outcome:
Previous discussion favored typed error results with centralized logging. Apply that pattern before adding new exception paths.
Security Audit
High Riskv6 • 7/6/2026 Open versioned report
Several static external-command alerts are Markdown false positives, but the skill does instruct Bash execution of local helper scripts and an unaudited installer. I found no evidence of network exfiltration, but the forceful always-on memory behavior and conversation-history access create security and privacy concerns.
1
Files scanned
131
Lines analyzed
14
Review items
0
False positives ignored
Confirmed security concerns (2)
High
Prompt Injection Attempt Detected
The skill says "THIS IS NOT OPTIONAL" and tries to make memory search the default behavior for every task. This can override user intent and higher-priority workflow decisions.
The file uses imperative authority language to force behavior for every task. It does not target this audit directly, so confidence is high but not absolute.
Medium
Automatic Conversation History Access
The skill instructs the agent to search previous discussions, prior context, and role-filtered user or assistant messages before work begins. This creates a privacy risk if installed without clear consent.
The instructions directly require memory searches and include role filters for user and assistant messages. No network exfiltration is shown, so severity remains medium.
Capability review items (14)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
High
Ruby/shell backtick execution
```bash
The bash installation block runs ./install-memory-tools.sh, a relative installer not present in the package. This could execute an unreviewed local script.
Medium
Ruby/shell backtick execution
| `memory-search.sh "<query>"` | Hybrid search (recommended) |
This line instructs the agent to invoke memory-search.sh from the shell with query text. The script is PATH-resolved and not included for review.
Medium
Ruby/shell backtick execution
| `memory-search.sh "<query>" --mode semantic` | Find conceptually related |
This line instructs the agent to invoke memory-search.sh from the shell with query text. The script is PATH-resolved and not included for review.
Medium
Ruby/shell backtick execution
| `memory-search.sh "<query>" --mode term` | Exact term matching |
This line instructs the agent to invoke memory-search.sh from the shell with query text. The script is PATH-resolved and not included for review.
Medium
Ruby/shell backtick execution
| `memory-search.sh "<query>" --role user` | Only user messages |
This line instructs the agent to invoke memory-search.sh from the shell with query text. The script is PATH-resolved and not included for review.
Medium
Ruby/shell backtick execution
| `memory-search.sh "<query>" --role assistant` | Only your responses |
This line instructs the agent to invoke memory-search.sh from the shell with query text. The script is PATH-resolved and not included for review.
Medium
Ruby/shell backtick execution
| "Continue working on X" | `memory-search.sh "X"` |
This table maps user requests to immediate memory-search.sh shell invocations. The query comes from user context and the external helper is not auditable in this package.
Medium
Ruby/shell backtick execution
| "Fix the issue we discussed" | `memory-search.sh "issue"`, `memory-search.sh "bug"` |
This table maps user requests to immediate memory-search.sh shell invocations. The query comes from user context and the external helper is not auditable in this package.
Medium
Ruby/shell backtick execution
| "Use the approach we agreed on" | `memory-search.sh "approach"`, `memory-search.sh "decision"` |
This table maps user requests to immediate memory-search.sh shell invocations. The query comes from user context and the external helper is not auditable in this package.
Medium
Ruby/shell backtick execution
| "Like we did before" | `memory-search.sh "<topic> implementation"` |
This table maps user requests to immediate memory-search.sh shell invocations. The query comes from user context and the external helper is not auditable in this package.
Medium
Ruby/shell backtick execution
| Any specific feature/component | `memory-search.sh "<feature>"` |
This table maps user requests to immediate memory-search.sh shell invocations. The query comes from user context and the external helper is not auditable in this package.
Medium
Ruby/shell backtick execution
| References to past work | `memory-search.sh "<reference>" --mode semantic` |
This table maps user requests to immediate memory-search.sh shell invocations. The query comes from user context and the external helper is not auditable in this package.
Medium
Ruby/shell backtick execution
```bash
The bash example block directs the agent to run memory-search.sh with several queries and role filters. The helper script is external to the package and cannot be audited here.
Medium
Ruby/shell backtick execution
```bash
The bash example combines memory-search.sh and doc-search.sh, both external commands. Their implementations are not included, so command behavior depends on the local environment.
Risk Factors
⚙️ External commands (26)
SKILL.md:33-38 SKILL.md:38-46 SKILL.md:46-47 SKILL.md:47-48 SKILL.md:48-49 SKILL.md:49-50 SKILL.md:50-56 SKILL.md:56-57 SKILL.md:57 SKILL.md:58 SKILL.md:59-60 SKILL.md:60-61 SKILL.md:61-65 SKILL.md:65-81 SKILL.md:81-87 SKILL.md:87-88 SKILL.md:88-89 SKILL.md:89-90 SKILL.md:90-105 SKILL.md:105-109 SKILL.md:109-119 SKILL.md:119-120 SKILL.md:120-121 SKILL.md:121-126 SKILL.md:126-128 SKILL.md:128-130
📁 Filesystem access (6)
Audited by: codex View Audit History →
Share & cite this report
Share the versioned assessment report, neutral badge, embed card, and citations. Skillstore reports evidence without deciding whether this Skill is safe.
Copy report link
https://skillstore.io/skills/23blocks-os-ai-maestro-memory-search/audits/6?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_reportMarkdown badge
[](https://skillstore.io/skills/23blocks-os-ai-maestro-memory-search?utm_source=security_passport_badge)HTML badge
<a href="https://skillstore.io/skills/23blocks-os-ai-maestro-memory-search?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/23blocks-os-ai-maestro-memory-search/security.svg" alt="Skillstore security assessment" loading="lazy"></a>Embed card
<iframe src="https://skillstore.io/embed/skills/23blocks-os-ai-maestro-memory-search.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>Academic citations (APA · BibTeX · CFF)
APA citation
23blocks-OS. (2026). ai-maestro-memory-search security audit report (audit version 6) [Author version unspecified]. Skillstore. https://skillstore.io/skills/23blocks-os-ai-maestro-memory-search/audits/6BibTeX citation
@techreport{23blocks-os-23blocks-os-ai-maestro-memory-search-2026,
author = {23blocks-OS},
title = {ai-maestro-memory-search security audit report (audit version 6)},
institution = {Skillstore},
year = {2026},
number = {6},
url = {https://skillstore.io/skills/23blocks-os-ai-maestro-memory-search/audits/6},
note = {Author version unspecified}
}CITATION.cff
cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "ai-maestro-memory-search security audit report (audit version 6)"
version: "unspecified"
type: report
authors:
- name: "23blocks-OS"
date-released: "2026-07-06"
url: "https://skillstore.io/skills/23blocks-os-ai-maestro-memory-search/audits/6"
identifiers:
- type: other
value: "skillstore:23blocks-os-ai-maestro-memory-search:audit:6"
description: "Skillstore immutable audit report identifier"
Skillstore Score
Why this score Evidence Confidence: Medium 55
Architecture
85
Maintainability
85
Content
71
Community
83
Spec Compliance
What You Can Build
Resume Development Work
Recover prior requirements, decisions, and partial implementation context before editing files.
Avoid Repeated Answers
Find previous explanations before responding, so the agent can build on existing context.
Review Prior Decisions
Locate earlier choices about architecture, bugs, or product behavior before planning new work.
Try These Prompts
Find Prior Context
Search memory for previous discussions about [topic], then summarize the relevant decisions before starting.
Continue Earlier Work
Search for past work on [feature], identify what remains, and explain the next implementation step.
Compare Memory And Docs
Search memory and documentation for [component], then list prior decisions, current guidance, and conflicts.
Investigate Repeated Issue
Search user and assistant history for [bug], group prior fixes, and recommend the safest next action.
Best Practices
- Search only for terms that are relevant to the current user request.
- Summarize found context before acting on it.
- Treat missing search results as useful context, not as a failure.
Avoid
- Do not run memory searches for unrelated or sensitive topics.
- Do not treat search results as current facts without verification.
- Do not install helper scripts without reviewing their source.
Frequently Asked Questions
What does this skill do?
It tells an AI agent to search prior conversation history before starting work.
Which tools does it support?
The report lists support for Claude, Codex, and Claude Code.
Does it include the search scripts?
No. The package contains instructions, but the helper scripts are not included.
Does it need shell access?
Yes. The documented workflows use Bash commands such as memory-search.sh.
Can it search only user or assistant messages?
Yes. The instructions describe role filters for user messages and assistant responses.
What should users verify before using it?
Users should verify the installed helper scripts, PATH target, and conversation data scope.
Developer Details
Author
23blocks-OSLicense
MIT
Skillstore revision
r1
Version notice
The author did not declare a version.
Ref
c1fdca50ff516318f65fed0d7f9e82797c5171dc
Maintenance freshness
7/18/2026
Usage
19 downloads · 236 views
File structure
📄 SKILL.md