MCP, MCP server
Ori Mnemos
Local-first persistent agentic memory powered by Recursive Memory Harness (RMH).
Compatibility: Not stated.
Grade B
Minor issues only, no Medium or High. Worth a quick look at the findings below before installing.
Not a guarantee. A hygiene check of the source at the commit shown, not a security audit.
Rescanned at commit 236ecaf4eed7, unchanged since the check below.
PERMISSION_DELETE_FILES NETWORK_NO_FIXED_HOST NETWORK_UNNAMED_HOST SECRETS_ENV_READ OBFUSCATION_FILE_OVER_READ_CAP
Why this grade
- PERMISSION_DELETE_FILES Low
Contains a call that can delete files.
adapters/hermes/plugin/hooks.py, line 37
result = subprocess.run( - NETWORK_NO_FIXED_HOST Low
Makes a network call whose host is not written as a fixed address.
bench/hotpotqa-eval.ts, line 365
const resp = await fetch("https://api.openai.com/v1/chat/completions", { - NETWORK_UNNAMED_HOST Low
Calls a fixed host that the description does not name.
bench/hotpotqa-eval.ts, line 365
const resp = await fetch("https://api.openai.com/v1/chat/completions", { - SECRETS_ENV_READ Note
Reads an environment variable whose name contains KEY, SECRET, TOKEN, PASSWORD or CREDENTIAL.
bench/hotpotqa-eval.ts, line 439
const apiKey = args.llmJudge ? (process.env.OPENAI_API_KEY ?? "") : ""; - OBFUSCATION_FILE_OVER_READ_CAP Note
A file was over the read cap and was not fully read.
bench/results/locomo-eval-2026-03-19T21-47-05-726Z.json, line 1
{
Any High, or two or more Medium findings, give D. One Medium gives C. Low findings or a security Note give B. Each code names a finding behind the letter.
Grades are a hygiene check of the source at the commit shown, on the date checked, not a security audit or a guarantee. The free directory grades are automated. Before release, a PlugVet engineer reviews every paid report and any grade we hold back, including all F grades, against the source code and the evidence we quote.
What was checked
Plain notes from the hygiene checks.
Permissions
Concern
can delete files.
PERMISSION_DELETE_FILES
adapters/hermes/plugin/hooks.py, line 37
result = subprocess.run(
adapters/hermes/plugin/hooks.py, line 103
subprocess.run(
bench/forgeteval_ori_adapter.py, line 34
self.proc = subprocess.Popen(
Show more
bench/hotpotqa-eval.ts, line 532
try { await fs.rm(vaultRoot, { recursive: true, force: true }); } catch {}
bench/locomo-eval.ts, line 762
try { await fs.rm(vaultRoot, { recursive: true, force: true }); } catch { /* best effort */ }
bench/longmemeval-eval.ts, line 214
await fs.rm(vault, { recursive: true, force: true });
bench/mem0-hotpotqa.py, line 264
shutil.rmtree(storage_dir, ignore_errors=True)
src/cli/bridge.ts, line 883
await fs.rm(pluginDir, { recursive: true, force: true });
src/cli/bridge.ts, line 1172
await fs.rm(pluginPath, { force: true });
src/cli/explore.ts, line 602
try { await fs.rm(sessionFilePath(vaultRoot, id)); } catch { /* already gone */ }
src/cli/init.ts, line 52
await fs.unlink(paths.marker);
docs/falsification/findings-contract.md, line 318
exit (every shell pipeline, `subprocess.run`, CI step) sees the timeout as ineffective. Filed as
Install and update
Pass
No install-time script found, and dependencies are pinned or absent.
Network
Concern
can reach api.openai.com. The description does not name that host. can reach a host that is not written as a fixed address.
NETWORK_NO_FIXED_HOST NETWORK_UNNAMED_HOST
bench/hotpotqa-eval.ts, line 365
const resp = await fetch("https://api.openai.com/v1/chat/completions", {
bench/locomo-eval.ts, line 529
const resp = await fetch("https://api.openai.com/v1/chat/completions", {
src/core/update-check.ts, line 103
const req = https.get(
Show more
src/providers/anthropic.ts, line 61
const response = await fetch(
src/providers/anthropic.ts, line 119
const response = await fetch(
src/providers/openai-compat.ts, line 65
const response = await fetch(
src/providers/openai-compat.ts, line 118
const response = await fetch(
Secrets
Note
reads an environment variable.
SECRETS_ENV_READ
bench/hotpotqa-eval.ts, line 439
const apiKey = args.llmJudge ? (process.env.OPENAI_API_KEY ?? "") : "";
bench/hotpotqa-eval.ts, line 455
const recApiKey = process.env.OPENAI_API_KEY ?? "";
bench/locomo-eval.ts, line 639
const apiKey = args.llmJudge ? process.env.OPENAI_API_KEY ?? "" : "";
Show more
bench/mem0-hotpotqa.py, line 184
api_key = os.environ.get("OPENAI_API_KEY", "")
Agent instructions
Pass
No instruction to ignore earlier instructions or to hide a step.
Obfuscation
Note
a file was over the read cap and was not fully read.
OBFUSCATION_FILE_OVER_READ_CAP
bench/results/locomo-eval-2026-03-19T21-47-05-726Z.json, line 1
{
bench/results/locomo-eval-2026-03-19T21-59-51-933Z.json, line 1
{
bench/results/locomo-eval-2026-03-19T22-09-11-676Z.json, line 1
{
Show more
bench/results/locomo-eval-2026-03-19T22-51-42-479Z.json, line 1
{
bench/results/locomo-eval-2026-03-19T23-07-39-929Z.json, line 1
{
bench/results/locomo-eval-2026-03-19T23-10-07-411Z.json, line 1
{
bench/results/locomo-eval-2026-03-20T01-48-17-281Z.json, line 1
{
bench/results/locomo-eval-2026-03-20T06-16-41-585Z.json, line 1
{
Licence
Pass
A licence file is present (LICENSE).
Maintainer activity
Pass
The latest commit is within the last 365 days.
Note
Hygiene check of commit 236ecaf4eed7. can delete files. can reach api.openai.com. The description does not name that host. can reach a host that is not written as a fixed address. reads an environment variable. a file was over the read cap and was not fully read.
Date and source
Date checked:
github.com
Commit graded: 236ecaf4eed766168cf8d1aae4f6d3aecbf683e5
This link goes to the original host. PlugVet does not host or check what you download from there, and installing it is your own decision and risk.