Codex, Plugin
Tmcra Memory
TMCRA Core — local-first, scope-isolated long-term memory runtime for AI agents.
Compatibility: Not stated.
Grade C
One Medium finding. Read the findings before installing.
Not a guarantee. A hygiene check of the source at the commit shown, not a security audit.
Rescanned at commit 72e5501742c5, unchanged since the check below.
SECRETS_PATH_CREDENTIAL
Why this grade
- SECRETS_PATH_CREDENTIAL Medium
Mentions a credential store path (~/.ssh, id_rsa, .npmrc, .aws/credentials, credentials.json, .netrc).
scripts/audit_public_release.py, line 184
if lower_name in {".env", "credentials.json", "secrets.json"}:
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
Note
requests shell access. The manifest names this permission.
PERMISSION_SHELL_DECLARED
benchmarks/longmemeval/src/tmcra_benchmark/legacy/run_tmcra_v4_build.py, line 55
result = subprocess.run(
benchmarks/longmemeval/src/tmcra_benchmark/legacy/run_tmcra_v4_build.py, line 118
subprocess.run(
benchmarks/longmemeval/src/tmcra_benchmark/legacy/run_tmcra_v4_evaluate.py, line 47
result = subprocess.run(
Show more
benchmarks/longmemeval/src/tmcra_benchmark/legacy/run_tmcra_v4_parallel_official_judge.py, line 115
completed = subprocess.run(
runtime/src/tmcra_local/__main__.py, line 257
subprocess.run(download, check=True, shell=False)
runtime/src/tmcra_local/__main__.py, line 258
subprocess.run(verify, check=True, shell=False)
runtime/src/tmcra_local/auth.py, line 16
completed = subprocess.run(
runtime/src/tmcra_local/auth.py, line 40
completed = subprocess.run(
Install and update
Pass
No install-time script found, and dependencies are pinned or absent.
Network
Concern
can reach a host that is not written as a fixed address.
NETWORK_NO_FIXED_HOST
benchmarks/longmemeval/src/tmcra_benchmark/harness/evaluate_qa_official_judge.py, line 143
with urllib.request.urlopen(req, timeout=timeout) as resp:
benchmarks/longmemeval/src/tmcra_benchmark/harness/run_lme_s10_native_tmcra.py, line 146
with urllib.request.urlopen(req, timeout=timeout) as response:
benchmarks/longmemeval/src/tmcra_benchmark/legacy/tmcra_v2_lme_pipeline.py, line 325
with urllib.request.urlopen(req, timeout=120) as resp:
Show more
benchmarks/longmemeval/src/tmcra_benchmark/legacy/tmcra_v3_product_writer.py, line 913
with urllib.request.urlopen(request, timeout=self.timeout) as response:
benchmarks/longmemeval/src/tmcra_benchmark/legacy/tmcra_v3_recall_planner.py, line 262
with urllib.request.urlopen(request, timeout=self.timeout) as response:
benchmarks/longmemeval/src/tmcra_benchmark/legacy/tmcra_v3_slow_graph.py, line 294
with urllib.request.urlopen(request, timeout=90) as response: # nosec B310
benchmarks/longmemeval/src/tmcra_benchmark/legacy/tmcra_v4_evidence_planner.py, line 495
with urllib.request.urlopen(request, timeout=self.timeout) as response:
benchmarks/longmemeval/src/tmcra_benchmark/legacy/tmcra_v4_recall_planner.py, line 244
with urllib.request.urlopen(request, timeout=self.timeout) as response:
Secrets
Concern
reads a credential or environment file.
SECRETS_ENV_READ SECRETS_PATH_CREDENTIAL SECRETS_PATH_DOTENV
benchmarks/longmemeval/src/tmcra_benchmark/harness/run_lme_s10_native_tmcra.py, line 284
answer_api_key = clean_text(os.getenv("TMCRA_ANSWER_API_KEY", ""))
benchmarks/longmemeval/src/tmcra_benchmark/harness/run_lme_s10_native_tmcra.py, line 289
answer_api_key or clean_text(os.getenv("OPENAI_API_KEY", "")),
scripts/audit_public_release.py, line 184
if lower_name in {".env", "credentials.json", "secrets.json"}:
Show more
benchmarks/longmemeval/src/tmcra_benchmark/harness/run_lme_s10_native_tmcra.py, line 294
clean_text(os.getenv("GEMMA_API_KEY", "")),
benchmarks/longmemeval/src/tmcra_benchmark/harness/run_lme_s10_native_tmcra.py, line 368
api_key=os.getenv("TMCRA_WRITER_API_KEY", ""),
benchmarks/longmemeval/src/tmcra_benchmark/harness/run_lme_s10_native_tmcra.py, line 371
max_tokens=int(os.getenv("TMCRA_WRITER_MAX_TOKENS", "512")),
benchmarks/longmemeval/src/tmcra_benchmark/harness/run_lme_s10_native_tmcra.py, line 714
clean_text(os.getenv("TMCRA_EVIDENCE_UNIT_PLANNER_API_KEY", "")) or answer_api_key,
benchmarks/longmemeval/src/tmcra_benchmark/harness/run_lme_s10_native_tmcra.py, line 724
clean_text(os.getenv("TMCRA_LLM_CHANNEL_PLANNER_API_KEY", "")) or answer_api_key,
benchmarks/longmemeval/src/tmcra_benchmark/harness/run_lme_s10_native_tmcra.py, line 2164
max_tokens = max(280, int(os.getenv("TMCRA_EVIDENCE_UNIT_PLANNER_MAX_TOKENS", "760") or 760))
.gitignore, line 7
# Local environments and provider credentials
.gitignore, line 8
.env
.gitignore, line 10
*.env
.gitignore, line 11
!*.env.example
.gitignore, line 17
credentials.json
.gitignore, line 28
*.key
SECURITY.md, line 33
Then review the exact staged diff and confirm that only paths permitted by `PUBLIC_RELEASE_MANIFEST.json` are included. Never publish `.tmcra/`, real `.env` files, key files, datab
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
benchmarks/longmemeval/src/tmcra_benchmark/legacy/audit_tmcra_v4_chain.py, line 1
"""Strict post-run audit for the TMCRA V4 writer/slow/retrieval chain.
benchmarks/longmemeval/src/tmcra_benchmark/legacy/tmcra_v4_batch_writer.py, line 1
from __future__ import annotations
benchmarks/longmemeval/src/tmcra_benchmark/legacy/tmcra_v4_slow_graph.py, line 1
#!/usr/bin/env python3
Show more
code/memory_adapters.py, line 1
from __future__ import annotations
runtime/src/tmcra_local/core/v4_batch_writer.py, line 1
from __future__ import annotations
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 72e5501742c5. requests shell access. The manifest names this permission. can reach a host that is not written as a fixed address. reads a credential or environment file. a file was over the read cap and was not fully read.
Date and source
Date checked:
github.com
Commit graded: 72e5501742c597a850ab5016aac798e71bcbab51
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.