Claude Code, Plugin
QAIA
🇫🇷 Lire ce README en français
Compatibility: Not stated.
Grade D
Two or more Medium findings, or a High. Only try it in a sandbox.
Not a guarantee. A hygiene check of the source at the commit shown, not a security audit.
AGENT_IGNORE_INSTRUCTIONS PERMISSION_SHELL_UNDECLARED
Why this grade
- AGENT_IGNORE_INSTRUCTIONS High
Text tells the agent to ignore earlier instructions.
plugins/qaia-core/skills/us-ingest/connectors/jira.md, line 46
directive (`ignore previous instructions`, a fake SYSTEM note, an embedded tool call) becomes a - PERMISSION_SHELL_UNDECLARED Medium
Runs shell commands that the manifest or description does not mention.
.claude/hooks/lint-skill-after-edit.py, line 46
run = subprocess.run([sys.executable, linter, path],
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.
If you think this grade is wrong, email [email protected] with the entry link and your evidence. We aim to acknowledge within two working days and decide within seven. If we confirm an error, we aim to correct it within 24 hours of confirming it. These are aims, not guarantees.
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
requests shell access. No manifest field names shell or command execution.
PERMISSION_SHELL_UNDECLARED
.claude/hooks/lint-skill-after-edit.py, line 46
run = subprocess.run([sys.executable, linter, path],
eval/tools/automation_score.py, line 971
p = subprocess.run(cmd, cwd=cwd, shell=False, timeout=timeout,
eval/tools/check_decision_register.py, line 59
out = subprocess.check_output(cmd, stderr=subprocess.STDOUT)
Show more
eval/tools/check_open_work_issue.py, line 86
out = subprocess.check_output(cmd, stderr=subprocess.STDOUT)
eval/tools/drift_campaign.py, line 152
proc = subprocess.run(
eval/tools/mutate_guards.py, line 205
p = subprocess.run(cmd, cwd=ROOT, capture_output=True, text=True, shell=False)
eval/tools/score_portability.py, line 33
p = subprocess.run(["npx", "--yes", "[email protected]", "-c", LINTRC, path],
eval/tools/selfcheck_gherkin_dialect.py, line 140
p = subprocess.run(cmd, capture_output=True, text=True)
Install and update
Concern
dependencies are not pinned.
INSTALL_UNPINNED
requirements.txt, line 8
PyYAML>=5.1
Network
Concern
can reach a host that is not written as a fixed address.
NETWORK_NO_FIXED_HOST
CLAUDE.md, line 24
curl -sL -H "Authorization: Bearer $GITHUB_PERSONAL_ACCESS_TOKEN" \
CLAUDE.md, line 28
curl -sL -H "Authorization: Bearer $GITHUB_PERSONAL_ACCESS_TOKEN" \
CLAUDE.md, line 32
curl -sL -H "Authorization: Bearer $GITHUB_PERSONAL_ACCESS_TOKEN" \
Show more
eval/contract-probe-2026-08-01/probe.sh, line 14
curl -s -X POST $B/api/reset -o /dev/null -w 'reset:%{http_code}\n'
eval/contract-probe-2026-08-01/probe.sh, line 17
curl -s -X POST $B/api/login -H 'Content-Type: application/json' \
eval/contract-probe-2026-08-01/probe.sh, line 35
out=$(curl -s -X "$m" "$B$p" -H "Authorization: Bearer $t" -H 'Content-Type: application/json' -d "$body" -w '\n<<%{http_code}>>')
eval/contract-probe-2026-08-01/probe.sh, line 37
out=$(curl -s -X "$m" "$B$p" -H "Authorization: Bearer $t" -w '\n<<%{http_code}>>')
eval/contract-probe-2026-08-01/probe.sh, line 47
id=$(curl -s -X POST $B/api/reports -H "Authorization: Bearer $tok" -H 'Content-Type: application/json' -d '{}' | jqid)
eval/uptime-kuma-2026-08-11/probe/p9-slug.js, line 36
const r = await fetch("http://127.0.0.1:3001/api/status-page/" + encodeURIComponent(slug));
Secrets
Concern
reads a credential or environment file.
SECRETS_ENV_READ SECRETS_PATH_DOTENV
eval/meilisearch-2026-08-11/probe.js, line 16
const KEY = process.env.MEILI_KEY || 'QAIAmasterKey2026probe';
.env.example, line 1
# Copy to .env and fill in real values. .env is gitignored - never commit real secrets.
.gitignore, line 1
# Secrets - never committed (T3)
Show more
.gitignore, line 2
.env
.gitignore, line 4
!.env.example
SECURITY.md, line 17
- **Generated code**: patterns in generated Playwright tests that could leak secrets (credentials in code, `.env` mishandling).
docs/DECISIONS.md, line 59
| D51 | fondateur | **Second juge LLM indépendant, outillage mainteneur uniquement** (`eval/tools/second_judge.py`) : chaîne de repli 100% gratuite (Gemini → Groq → Hugging Face),
docs/DECISIONS.md, line 149
| T3 | Q52 | Secrets/environnements : **jamais dans la session ni le RAG** ; pattern `.env` + fixtures Playwright. |
docs/DISCOVERY.md, line 131
52. **Secrets & environnements** : URL, comptes de test, credentials - jamais dans la session ni dans le RAG versionné. Quel pattern imposé dès la conception (`.env`, vault, fixtur
Agent instructions
Serious
instructions tell the agent to ignore earlier instructions.
AGENT_IGNORE_INSTRUCTIONS
plugins/qaia-core/skills/us-ingest/connectors/jira.md, line 46
directive (`ignore previous instructions`, a fake SYSTEM note, an embedded tool call) becomes a
Obfuscation
Note
a file was over the read cap and was not fully read. the file cap was reached before every file was read.
OBFUSCATION_FILE_OVER_READ_CAP OBFUSCATION_FILE_CAP_REACHED
docs/DECISIONS.md, line 1
# Décisions de discovery - QAIA
eval/architecture-review-2026-08-08/raw.json, line 1
{
eval/openapi-ingest-2026-08-08/testbook/petstore-pets.feature, line 1
# Test book -- Swagger Petstore, /pet operations.
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 b08095b48edd. requests shell access. No manifest field names shell or command execution. dependencies are not pinned. can reach a host that is not written as a fixed address. reads a credential or environment file. instructions tell the agent to ignore earlier instructions. a file was over the read cap and was not fully read. the file cap was reached before every file was read.
Date and source
Date checked:
github.com
Commit graded: b08095b48edd462f958f37762bb3c22b77979314
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.