memory-system-v2
Verdict safe, but 1 finding (1 medium), mainly data exfiltration. Most recent scan 2026-02-06.
The memory-system-v2 skill is a legitimate local memory management system with LOW to MEDIUM security concerns. The skill's primary function is to store and retrieve memory entries in the user's home directory using bash and jq. No critical threats were identified - there is no data exfiltration, no network calls, no credential theft, no command injection, and no prompt injection attempts. The main concern is the lack of declared tool restrictions (allowed-tools field), which is a transparency issue rather than an active threat. The skill operates entirely locally, creates files in a dedicated directory structure, and performs standard file operations. The code is straightforward bash scripting without obfuscation or malicious patterns. Missing referenced files are documentation issues only. Overall, this skill appears safe for use with minor improvements recommended for transparency and resource management.
Scanner's overall assessment (LLM analyzer)
Second scanner: NVIDIA SkillSpector
Independent static run, no language model, over the same files as the Cisco scan above. How the two scanners compare across the whole corpus: Two scanners, one corpus.
What this skill does
Fast semantic memory system with JSON indexing, auto-consolidation, and <20ms search. Capture learnings, decisions, insights, events. Use when you need persistent memory across sessions or want to recall prior work/decisions.
- Package
- 5 files, 29 KB
- SKILL.md
- 1153 words, 12+ sections
- Code
- Shell (1)
- Requires
- bins:
jq
SKILL.md covers: Memory System v2.0 · Overview · Features · Quick Start · Installation · Install jq (required dependency) · Copy memory-cli.sh to your workspace · Already installed if you're using Clawdbot · Basic Usage · Memory Types · 1. Learning (importance: 7-9) · 2. Decision (importance: 6-9)
Findings
MEDIUM Unrestricted File System Access Without Tool Restrictions Data Exfiltrationllm
The skill creates and manages files in $HOME/clawd/memory directory without declaring allowed-tools in the manifest. While the operations appear benign (memory storage), the absence of tool restrictions means there's no enforcement mechanism to prevent the skill from performing unauthorized file operations. The skill reads/writes to user's home directory without explicit permission declaration.
memory-cli.sh
MEMORY_DIR="$HOME/clawd/memory"
mkdir -p "$INDEX_DIR" "$DAILY_DIR" "$CONSOLIDATED_DIR"
echo '...' > "$INDEX_DIR/memory-index.json"What is in the package
| File | Kind | Size |
|---|---|---|
README.md | doc | 908 B |
SKILL.md | doc | 9.1 KB |
memory-cli.sh | script | 6.6 KB |
docs/memory-system-v2-design.md | doc | 4.3 KB |
docs/memory-system-v2-test-results.md | doc | 8.0 KB |
External hosts referenced
github.com
Hostnames that appear in the skill's files. Referencing a host is not a finding; compare against the findings above and the skill's stated purpose.
Environment variables referenced
INDEX_DIR MEMORY_DIR CONSOLIDATED_DIR DAILY_DIR
Package contents recorded 2026-10-22. The scan above ran against the version current at scan time; the two can differ if the skill has since been updated.
What we checked
- Static analyzer. Pattern rules over SKILL.md and bundled files: known-bad commands, exfiltration URLs, secrets, obfuscation.
- Behavioral analyzer. Checks what the skill asks the agent to do against what its description claims it does.
- Llm analyzer. A language model reads the skill and reports intent-level issues the rules cannot express.
- Trigger analyzer. Looks for instructions that fire on conditions the user did not ask for.
- Meta analyzer. A second model pass that validates, deduplicates and re-scores the other analyzers' findings.
What we did not check
- Runtime behaviour. The skill was not executed; findings come from reading its files.
- Anything fetched at install or run time, including remote scripts, package dependencies, and URLs the skill tells the agent to open.
- Content the analyzers cannot read: images, binaries, and some non-English text.
- Known scanner bypass classes. See the methodology and limitations page.
Sources
- Repository: https://clawhub.ai/skills/memory-system-v2
- Raw scan result: memory-system-v2-scan.json
- Badge for your README:
- Request a re-scan