owlmem@0.9.15
OwlMem command line: connect your AI apps to OwlMem in one command (npx -y owlmem install), an end-to-end encrypted vault for your AI agents, and a local MCP server.
AI Security Review
scanned 8h ago · by lpm-firewall-aiThe flagged credential flow reads Claude Code’s sign-in only for an owner-enabled quota feature and sends it to Anthropic’s own usage endpoint; resulting quota data goes to OwlMem. The CLI’s execution and background features require explicit commands or setup, and no unrelated collection or hidden automatic behavior was found.
Static reason
High-risk behavior combination matched malicious policy.
Trigger
CLI commands run when invoked; auto-wake requires explicit setup, and Claude quota reading requires an owner-enabled setting.
Impact
The reviewed behavior matches the package’s stated features; no unrelated credential collection or exfiltration was identified.
Mechanism
Vault requests use the configured OwlMem service. When enabled, Claude quota lookup sends Claude Code’s access token to api.anthropic.com and reports quota data to OwlMem.
AI rationale
The flagged credential flow reads Claude Code’s sign-in only for an owner-enabled quota feature and sends it to Anthropic’s own usage endpoint; resulting quota data goes to OwlMem. The CLI’s execution and background features require explicit commands or setup, and no unrelated collection or hidden automatic behavior was found.
Evidence
package.jsonowlmem.mjs
Network endpoints3
owlmem.comapi.anthropic.com/api/oauth/usageregistry.npmjs.org
Decision evidence
public snapshotAI called this Clean at 94.0% confidence as Benign with low false-positive risk.
Evidence for AI clean decision
- The manifest defines a CLI entrypoint and has no lifecycle scripts.
- The CLI sends its configured OwlMem token to the OwlMem API base URL for vault requests.
- Claude credentials are read only for a quota feature described as requiring the owner to enable it; the token is sent to Anthropic’s own usage endpoint.
- The quota is fetched only when the saved OwlMem configuration has `read_quota` enabled for Claude Code.
- The command-line handler is guarded, so importing the module does not automatically run the CLI.
Evidence against
- The CLI can execute a user-specified command and pass it requested vault secrets via `owlmem run`.
- The explicitly enabled auto-wake feature runs in the background and reports usage status to OwlMem.
Behavioral surface
ChildProcessCryptoEnvironmentVarsFilesystemNetworkShell
HighEntropyStringsUrlStrings
Source & flagged code
10 flaggedowlmem.mjsView file
16// owlmem vault ls
L17: // owlmem vault get <name> [field] prints the item (or one field) to stdout
L18: // owlmem vault put <name> [--desc text] < value-or-json
...
L39: import { readFileSync, writeFileSync, mkdirSync, existsSync, chmodSync, copyFileSync, rmSync, statSync, realpathSync, openSync, closeSync, appendFileSync, renameSync, readdirSync, ...
L40: import { gunzipSync } from "node:zlib";
L41: import { homedir, hostname, userInfo } from "node:os";
...
L43: import { fileURLToPath } from "node:url";
L44: import { spawn, spawnSync } from "node:child_process";
L45:
L46: export const VERSION = "0.9.15";
L47: const BASE = (process.env.OWLMEM_URL || "https://owlmem.com").replace(/\/$/, "");
L48: const DIR = process.env.OWLMEM_HOME || join(process.env.XDG_CONFIG_HOME || jo
Critical
Credential Exfiltration
Source appears to send environment or credential material to an external endpoint.
owlmem.mjsView on unpkg · L1616Trigger-reachable chain: manifest.bin -> owlmem.mjs
L16: // owlmem vault ls
L17: // owlmem vault get <name> [field] prints the item (or one field) to stdout
L18: // owlmem vault put <name> [--desc text] < value-or-json
...
L39: import { readFileSync, writeFileSync, mkdirSync, existsSync, chmodSync, copyFileSync, rmSync, statSync, realpathSync, openSync, closeSync, appendFileSync, renameSync, readdirSync, ...
L40: import { gunzipSync } from "node:zlib";
L41: import { homedir, hostname, userInfo } from "node:os";
...
L43: import { fileURLToPath } from "node:url";
L44: import { spawn, spawnSync } from "node:child_process";
L45:
L46: export const VERSION = "0.9.15";
L47: const BASE = (process.env.OWLMEM_URL || "https://owlmem.com").replace(/\/$/, "");
L48: const DIR = process.env.O
Findings
2 Critical8 High3 Medium3 Low
CriticalCredential Exfiltrationowlmem.mjs
CriticalTrigger Reachable Dangerous Capabilityowlmem.mjs
HighChild Processowlmem.mjs
HighShellowlmem.mjs
HighSame File Env Network Executionowlmem.mjs
HighCommand Output Exfiltrationowlmem.mjs
HighSandbox Evasion Gated Capabilityowlmem.mjs
HighTrigger Reachable Credential Exfiltrationowlmem.mjs
HighTrigger Reachable Command Output Exfiltrationowlmem.mjs
HighSemantic Analysis Limitedowlmem.mjs
MediumNetwork
MediumEnvironment Vars
MediumStructural Risk Force Deep Review
Affected versions and remediation
This report applies to owlmem@0.9.15.
See all versions of owlmem for other recorded verdicts.
- Review the evidence and your use of owlmem@0.9.15 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.