SKILLEMALL.ai
GitHub ↗ RU Sign in

FC gstack

Router for the gstack skill suite. (gstack)

The skillemall take

A router for the gstack skill suite that promises to wire up multiple tools into one system. The review found six high-severity data exfiltration issues. The code contains mechanisms that can send information outside the sandbox. Code quality scores 62, process score 64—both unremarkable. Scripts stayed quiet in sandbox, but that doesn't erase what's in the files.

The F grade and "not recommended" verdict aren't arbitrary. Six high-level exfiltration findings mean something. Don't install until the author fixes the data export problem.

Not recommendedcritical or high security findings · low grade F
garrytan/gstack Agent Skills author: garrytan MIT 80 files · 29 scripts body ≈ 4 009 tokens Open the sourcegithub.com↗ analyzed 28 h ago

Router for the gstack skill suite.

As a process C 64/100 · Has gaps — weak spots: result and completion, when it triggers, inputs and preconditions

ProcedureAI and agentstype and topics are labelled automatically from the skill text
Runs in: Claude / Claude Code ChatGPT / Codex Gemini CLI Cursor GitHub Copilot / VS Code Qwen Code Kimi Code CLI Z.ai (GLM) DeepSeek (Deep Code) Mistral Vibe LangChain (deepagents) OpenClaw Goose OpenCode
JSON
Technical rating
F
31/100
safety, quality, tests
Safety 60%
11
Quality 40%
62
Run on models
none yet
Process rating
C
64/100
Has gaps
Result and completion w 14
0
When it triggers w 12
20
Inputs and preconditions w 11
30
the three weakest of ten parameters · all ten

What is at stake

The skill contains fragments that, in the wrong hands, cost money or data. Below: what the installer risks and what the author should do.

Exfiltration
If you install

The instructions or scripts send data out: environment variables, keys, file contents, chat history. You may never notice, because the agent performs the upload as "part of the task".

For the author

If the upload is not needed for the task, remove it: catalogs flag such skills and delist them. If it is needed, name the destination explicitly, say what leaves the machine, and give the user a switch.

Dangerous commands medium severity

Below is the worst case for this category. The finding here is medium: the guard saw a sign, not a proof.

If you install

The skill contains commands that delete files, rewrite disks or execute code fetched from the network. The agent may run them without asking if it believes the instructions require it.

For the author

Replace destructive commands with safe equivalents that ask for confirmation, scope them to one folder, and stop piping curl into a shell: pin a version and a checksum.

Broad scope medium severity

Below is the worst case for this category. The finding here is medium: the guard saw a sign, not a proof.

If you install

The skill asks for more than the task needs: broad tool access, credential environment variables, binaries. Every extra permission widens the damage from a mistake or a compromise.

For the author

Narrow allowed-tools and the variable list to the minimum; replace binaries with readable sources or scripts.

How to improve

  1. Address the high-severity findings: each costs 18 safety points. If one is a false positive, add the rule id to guard.allow in spec.yaml.
  2. Say in the description WHEN to use the skill ("use when…", example requests): that is the agent's main cue.
For the model run — optional
  • Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
  • A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.

Guard findings · 9

  • high Exfiltration intent-browser-credential-store ARCHITECTURE.md:29
    Accesses a browser credential / cookie store
    The second engine is the one gstack has always shipped, and it is what runs everywhere Aside does not. An AI agent driving a browser it owns needs **sub-second latency** and **persistent state**: if e
  • high Exfiltration intent-browser-credential-store ARCHITECTURE.md:184
    Accesses a browser credential / cookie store
    1. **OS key access follows platform permissions.** macOS may prompt for Keychain approval on the first import per browser. Linux supports libsecret-backed `v11` and the Chromium fallback key for `v10`
  • high Exfiltration intent-browser-credential-store bin/gstack-doctor-components.sh:65
    Accesses a browser credential / cookie store
    browse|qa|qa-only|design-review|design-consultation|design-html|devex-review|scrape|skillify|canary|benchmark|make-pdf|diagram|pair-agent|open-gstack-browser|setup-browser-cookies|connect-chrome|land-
  • high Exfiltration intent-browser-credential-store bin/gstack-doctor-components.sh:78
    Accesses a browser credential / cookie store
    browse|qa|qa-only|design-review|design-consultation|design-html|devex-review|scrape|skillify|canary|benchmark|make-pdf|diagram|pair-agent|open-gstack-browser|setup-browser-cookies|connect-chrome|land-
Medium and low: 5
  • medium Dangerous commands cmd-execpolicy-bypass .github/workflows/windows-setup-e2e.yml:228
    Runs PowerShell with execution policy bypassed
    powershell.exe -NoProfile -NonInteractive -ExecutionPolicy Bypass -File "$(cygpath -w "$SPACED/setup.ps1")" --host codex 2>&1)
  • medium Exfiltration intent-browser-credential-store bin/gstack-doctor-components.sh:11
    Accesses a browser credential / cookie store (quoted — discussed, not commanded)
    GSTACK_DOCTOR_SKILLS="autoplan plan-ceo-review plan-eng-review plan-design-review plan-devex-review plan-tune office-hours spec review ship investigate codex cso browse qa qa-only design-review design
    quoted
  • medium Broad scope meta-broad-allowed-tools SKILL.md:1
    Broad tool permissions pre-approved: Bash
    allowed-tools: Bash Read AskUserQuestion
  • low Secrets in code secret-password-literal .github/scripts/run-dia-native-qualification.ts:413
    Hard-coded password / key literal (may be an example)
    const password = randomBytes(24).toString('hex');
  • low Dangerous commands cmd-execpolicy-bypass .github/workflows/windows-setup-e2e.yml:453
    Runs PowerShell with execution policy bypassed (string literal in code, not executed)
    "$PS" -NoProfile -NonInteractive -ExecutionPolicy Bypass -File "$(w "$T/form.ps1")" 2>&1
    code literal

Files scanned: 78. Evidence is masked. Grey chips explain why severity was lowered.

Against the Agent Skills spec

  • warning description-no-when description does not say WHEN to use the skill (no "use when")
  • note frontmatter-key unknown frontmatter key "preamble-tier"
  • note frontmatter-key unknown frontmatter key "triggers"

Process rating: all ten parameters 64/100

  • 0Result and completion. Does not say what the result is
  • 20When it triggers. No condition that starts the skill
  • 30Inputs and preconditions. Does not say what the process needs to start
  • 30Running it twice. 12 mutating operations with no state check
  • 70Execution cost. Instruction body is 4009 tokens
  • 100Tools and files. Tools declared in frontmatter
  • 100Steps. 47 steps
  • 100Failures and branches. 8 branches, has a failure section
  • 100Consistency. Name and required fields are in place
  • 100Progress reporting. Reports progress
  • medium Safety rules and hard prohibitions inside a skill: they belong in the system prompt, here they protect nothing
  • low 12 top-level sections: this looks like several domains in one skill

Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.

Quality signals

  • +5Description has no quoted example phrases that should trigger the skill
  • +4Description does not say when NOT to use the skill (false activations)
  • +3Description length 43: 120–800 characters recommended
  • +3Output format is not stated: the model decides each time
  • -5TODO / placeholder text left in the skill
  • +1No license
  • +2Single-language instructions
  • +4Structure: 12 headings
  • +3Step-by-step instructions: 47 items
  • +4Has examples (4 code blocks)

Quality base 70; lint remarks subtract, signals add up to 100. Result: 62.

In the sandbox The scripts kept to themselves

The skill's scripts were run in a throwaway machine: no network, fake keys in the home directory, a tracer watching. We wrote down what they did. Reaching for the network or for secrets caps the technical grade at C; a quiet run adds no points.

Запущено 12 скриптов, ещё 16 не поместились в лимит; каждому дали двадцать секунд, поддельный домашний каталог с ключами и сеть, в которой ничего нет.

Из них 2 не дошли до работы, и об их поведении мы ничего не узнали.

bin/gstack-autoplan-snapshot.tsничего за пределами своей папки
bin/gstack-brain-context-load.tsничего за пределами своей папки
bin/gstack-brain-lock.shничего за пределами своей папки
bin/gstack-brain-sync-index.pyне запустился: BRAIN_SYNC: staged artifact validation refused: list index out of range
bin/gstack-bun-version.shничего за пределами своей папки
bin/gstack-claude-root.shничего за пределами своей папки
bin/gstack-codex-status.shничего за пределами своей папки
bin/gstack-design-detect.tsничего за пределами своей папки
bin/gstack-design-md.tsничего за пределами своей папки
bin/gstack-doctor-check.shне запустился: bin/gstack-doctor-check.sh: line 22: /gstack-doctor-components.sh: No such file or directory
bin/gstack-doctor-components.shничего за пределами своей папки
bin/gstack-egress-lib.shничего за пределами своей папки

19 Oct 2026