Skills/Glama/ai-memory

ai-memory@r1

CAUTION· install only

Publisher claims: "Gives AI agents a shared, encrypted memory store accessible over MCP, with per-agent keys, per-namespace permissions, history and revert, attachments, and semantic search that handles Czech as well as English. Memories, agents, network access, and HTTPS can be managed through a self-hosted admin web on the same port, with backups being simply copies of the encrypted data file." Observed in sandbox: no network, touches nothing outside the project, writes no memory.

Registry GlamaPublisher tursithepanda · publishing for 1 daysInstalls 0Last detonated 19 h ago
What would fix thisWatchDispute verdict

Coverage

Nothing executable in the bundle ran; the trace covers the install step only.
Tested
every file read, file written and host contacted by what ran
Not tested
use of the skill by an agent

What this means for the label: Caution here covers what ran. The server's tools were not exercised with a real credential, so nothing about their behavior is asserted; the credential summary below says what to do with that.

Before you give it your token

It will ask for
no credential variables read in the code
Where it talks to
no vendor host named in the codeAlso huggingface.co — not the publisher's domain; check what for before connecting.
What the honeypot saw
no credential files read · no network
What we could not test
use of the skill by an agent

Connect it under the guard. Nothing executable in the bundle ran; the trace covers the install step only. The code names no credential variable; the guard pins the network to what the honeypot saw and logs every call.

$vet-skills run --net declared -- <the server's command>

Never paste a credential into a chat, a ticket or a report, including ours. How the guard holds it.

Is ai-memory@r1 safe to install?

KillerSkills ran it in the KillerSkills honeypot sandbox (1 scenario; nothing executable in the bundle ran; the trace covers the install step only) on 2026-11-13 and scored it 68/100 (Caution). It is not known to be malicious, but it has behaviour or code a security policy should decide on before install. Notable findings: Privileged sudo command; Privileged sudo command; README directs the reader to a committed installer. Summary: no network, touches nothing outside the project, writes no memory.

Behavior trace

Honeypot run · 0 s · 1 scenarios · 0 deviations
t+CategoryObserved action
0.0sprocessno install scripts in bundle
0.0sprocessscript backend runs setup only
0.0sprocessscript backend runs setup only
0.0sprocessno install scripts to re-run

Static findings

HIGH
Privileged sudo commandai_memory/server/netconf.py:293 · Privileged sudo command — agent may execute with elevated privilegesreturn (f"sudo firewall-cmd --add-port={port}/tcp --permanent && sudo firewall-cmd --reload"Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
Python subprocess execution detectedai_memory/cli.py:134 · Python subprocess execution detectedout = subprocess.run(["tasklist", "/FI", f"PID eq {pid}", "/NH"],Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05
MEDIUM
File permission modification commandDockerfile:90 · File permission modification command — agent changes file ownership/permissions&& chown -R 10001:10001 /data /backup \Fix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts.
AST05
MEDIUM
File permission modification commandDockerfile:91 · File permission modification command — agent changes file ownership/permissions&& chmod 0700 /dataFix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts.
AST05
HIGH
Privileged sudo commandai_memory/server/netconf.py:295 · Privileged sudo command — agent may execute with elevated privilegesf"sudo ufw allow {port}/tcp"Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
File permission modification commandDockerfile:85 · File permission modification command — agent changes file ownership/permissions# `chown 10001:10001`. No login shell, no sudo.Fix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts.
AST05
MEDIUM
File permission modification commandDockerfile:122 · File permission modification command — agent changes file ownership/permissions# chown above happens BEFORE this line: a fresh named volume is initialised from theFix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts.
AST05
CRITICAL
README directs the reader to a committed installerREADME.md:32**Windows:** download `ailurus-memory-<version>-windows-setup.exe` (or the portable zip)Fix: Pin versions and sources; vendor anything fetched at install time.
AST06
MEDIUM
Python subprocess execution detectedai_memory/cli.py:149 · Python subprocess execution detectedsubprocess.run(["taskkill", "/PID", str(pid), "/T", "/F"],Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05
CRITICAL
README tells the reader to bypass SmartScreenREADME.md:35SmartScreen may warn about an unsigned program: *More info → Run anyway*.Fix: Pin versions and sources; vendor anything fetched at install time.
AST06
MEDIUM
File permission modification commandDockerfile:141 · File permission modification command — agent changes file ownership/permissionsRUN chown -R 10001:10001 /app/testsFix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts.
AST05
MEDIUM
Python subprocess execution detectedai_memory/cli.py:360 · Python subprocess execution detectedprocs.append((label, subprocess.Popen(Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05
CRITICAL
README tells the reader to click past SmartScreenREADME.md:35SmartScreen may warn about an unsigned program: *More info → Run anyway*.Fix: Pin versions and sources; vendor anything fetched at install time.
AST06
MEDIUM
Python subprocess execution detectedai_memory/cli.py:264 · Python subprocess execution detectedproc = subprocess.Popen(Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05
LOW
Prompt injectiondocs/install/windows.md:89 · Prompt injection — known jailbreak technique reference · informational: not an instruction fileApps → Advanced settings → Developer mode, on Plus and above — but only for servers it canFix: Keep SKILL.md to visible, descriptive text. Remove comments and encoded text that address the agent.
AST01
MEDIUM
Python subprocess execution detectedtests/test_supervisor.py:31 · Python subprocess execution detectedreturn subprocess.Popen([sys.executable, "-c", code])Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05
MEDIUM
Python subprocess execution detectedtests/test_index.py:180 · Python subprocess execution detectedproc = subprocess.Popen([sys.executable, "-m", "ai_memory.server.server"], env=env,Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05
MEDIUM
Python subprocess execution detectedai_memory/core/unlock.py:190 · Python subprocess execution detectedout = subprocess.run([exe, "--timeout=300", prompt],Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05
MEDIUM
File permission modification commandpackaging/make-portable.sh:58 · File permission modification command — agent changes file ownership/permissionschmod 755 "$D/ai-memory"Fix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts.
AST05
MEDIUM
Privileged sudo commanddocker-compose.yml:74 · Privileged sudo command — agent may execute with elevated privileges# # sudo chown 10001:10001 ai_memory_passphrase.txt && sudo chmod 400 ai_memory_passphrase.txtFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
Python subprocess execution detectedtests/test_install.py:37 · Python subprocess execution detectedr = subprocess.run(["bash", "-c", "echo ok"], capture_output=True, text=True,Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05
MEDIUM
Python subprocess execution detectedtests/test_install.py:117 · Python subprocess execution detectedr = subprocess.run(["bash", "-n", str(sh)], capture_output=True, text=True)Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05
MEDIUM
Python subprocess execution detectedtests/test_install.py:121 · Python subprocess execution detectedr = subprocess.run(["bash", str(sh), "--nonsense"], capture_output=True, text=True)Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05
MEDIUM
Python subprocess execution detectedtests/test_install.py:119 · Python subprocess execution detectedr = subprocess.run(["bash", str(sh), "--help"], capture_output=True, text=True)Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05
MEDIUM
Python subprocess execution detectedtests/test_http.py:139 · Python subprocess execution detectedproc = subprocess.Popen([sys.executable, "-m", "ai_memory.server.server"], env=env,Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05
MEDIUM
Python subprocess execution detectedtests/test_install.py:102 · Python subprocess execution detectedr = subprocess.run([sys.executable, "-m", "ai_memory.cli", "install-service", "--print"],Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05
MEDIUM
Python subprocess execution detectedtests/test_indicator.py:94 · Python subprocess execution detectedr = subprocess.run(["/usr/bin/python3", "-c",Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05
MEDIUM
Python subprocess execution detectedai_memory/desktop/indicator.py:120 · Python subprocess execution detectedweb.connect("activate", lambda *_: subprocess.Popen(Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05
MEDIUM
Python subprocess execution detectedai_memory/desktop/indicator.py:81 · Python subprocess execution detectedsubprocess.run(["systemctl", "--user", *args, UNIT], check=False)Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05
MEDIUM
Process kill commandai_memory/cli.py:575 · Process kill command — agent can terminate processes# in a loop would throw dialogs at the user until they get it right or kill the unit.Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
Python subprocess execution detectedtests/test_wizard.py:35 · Python subprocess execution detectedreturn subprocess.run([sys.executable, "-m", "ai_memory.cli", *args],Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05
MEDIUM
Python subprocess execution detectedai_memory/desktop/tray.py:59 · Python subprocess execution detectedself.procs.append((label, subprocess.Popen(Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05
MEDIUM
Python subprocess execution detectedai_memory/core/unlock.py:112 · Python subprocess execution detectedout = subprocess.run([path, *args], capture_output=True, text=True,Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05
MEDIUM
File permission modification commanddocker-compose.yml:74 · File permission modification command — agent changes file ownership/permissions# # sudo chown 10001:10001 ai_memory_passphrase.txt && sudo chmod 400 ai_memory_passphrase.txtFix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts.
AST05
LOW
Authorization escalation languagedocs/install/windows.md:89 · Authorization escalation language · informational: not an instruction fileApps → Advanced settings → Developer mode, on Plus and above — but only for servers it canFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.
AST01
MEDIUM
Process kill commandai_memory/features/export.py:250 · Process kill command — agent can terminate processesexcept Exception as exc: # a broken export must never kill the webFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
Python subprocess execution detectedai_memory/desktop/tray.py:101 · Python subprocess execution detectedsubprocess.Popen(["xdg-open", str(path)])Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05
MEDIUM
Python subprocess execution detectedai_memory/cli.py:433 · Python subprocess execution detectedsubprocess.Popen(child_command_self(args), creationflags=flags,Fix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05
MEDIUM
Privileged sudo commanddocker-compose.yml:61 · Privileged sudo command — agent may execute with elevated privileges# mount instead, prepare it first: mkdir -p ./data && sudo chown 10001:10001 ./dataFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
File permission modification commanddocker-compose.yml:61 · File permission modification command — agent changes file ownership/permissions# mount instead, prepare it first: mkdir -p ./data && sudo chown 10001:10001 ./dataFix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts.
AST05

Cost to the agent

Instructions loaded
≈38 tok
1 file · 81 in bundle

Tokens estimated at four characters each. Skills load their instruction text when triggered; the figure is what that costs.

Known vulnerabilities

Vulnerability advisories (OSV) cover npm and PyPI packages; this registry has no advisory database.

What would make this Verified

12 points to 80 · weights: behavior 45%, static 25%, prompt 15%, publisher 15%
  1. README directs the reader to a committed installer≈ +9 on the totalPin versions and sources; vendor anything fetched at install time.
  2. README tells the reader to bypass SmartScreen≈ +9 on the totalPin versions and sources; vendor anything fetched at install time.
  3. README tells the reader to click past SmartScreen≈ +9 on the totalPin versions and sources; vendor anything fetched at install time.
  4. Unclaimed publisher≈ +8 on the totalClaim tursithepanda: a verified publisher gets +50 on publisher trust. It is free and usually instant with GitHub sign-in.
  5. Privileged sudo command≈ +4 on the totalBuild commands from fixed arguments; do not interpolate untrusted strings into a shell.
  6. Python subprocess execution detected≈ +2 on the totalBuild commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
Publish a new version to Glama, then queue it here. Not your skill? Claim the handle to get verdict changes by email.Claim tursithepandaPropose a patch to the maintainer