building-inferencesh-apps
8 findings (2 high, 2 medium, 3 low, 1 info), mainly data exfiltration, command injection, policy violation. Most recent scan 2026-10-22.
This is a byte-identical copy of building-inferencesh-apps from 101-skills, which is the primary page for this skill. Same SKILL.md, same files; each copy is scanned separately so divergence shows up here.
Second scanner: NVIDIA SkillSpector
Independent static run, no language model, over the same files as the Cisco scan above. How the two scanners compare across the whole corpus: Two scanners, one corpus.
What this skill does
Build and deploy applications on inference.sh. Use when getting started, understanding the platform, creating apps, configuring resources, or needing an overview of inference.sh app development. Supports both Python and Node.js. Triggers: inference.sh app, belt app, inf.yml, inference.py, inference.js, deploy app, app development, build app, create app, GPU app, VRAM, app resources, app secrets, app integrations, multi-function app
- Package
- 10 files, 48 KB
- SKILL.md
- 1422 words, 12+ sections
- Code
- None. Instructions only.
SKILL.md covers: Inference.sh App Development · Rules · CLI Installation · Quick Start · Development Workflow (mandatory) · 1. Scaffold · 2. Implement · 3. Test Locally · 4. Deploy · 5. Cloud Test & Verify · Other useful commands · App Structure
Findings
HIGH Dangerous data flow in command pipeline Data Exfiltrationpipeline
Pipeline downloads data from the network and executes it: `curl -fsSL https://cli.inference.sh | sh`. This is a remote code execution pattern.
SKILL.md:1
curl -fsSL https://cli.inference.sh | shHIGH Dangerous data flow in command pipeline Data Exfiltrationpipeline
Pipeline downloads data from the network and executes it: `curl -LsSf https://astral.sh/uv/install.sh | sh`. This is a remote code execution pattern. (Note: source host is listed as an installer endpoint, but artifact integrity was not verified.)
references/cli.md:2
curl -LsSf https://astral.sh/uv/install.sh | shMEDIUM Network-derived content reaches code execution Command Injectioncorrelation
Correlated network → code_execution behavior across SKILL.md.
SKILL.md:24
MEDIUM Remote fetch followed by execution Command Injectionpipeline
Content is downloaded from the network and subsequently executed. This is a classic remote code execution attack pattern. (found in documentation — may be instructional) (installer endpoint configured; artifact integrity not verified)
references/cli.md:8
# macOS / Linux
curl -LsSf https://astral.sh/uv/install.sh | sh
# Windows
powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"
LOW Dangerous data flow in command pipeline Data Exfiltrationpipeline
Pipeline downloads data from the network and executes it: `curl -fsSL https://cli.inference.sh | sh`. This is a remote code execution pattern. (Note: found in documentation file - may be instructional rather than executable.)
references/cli.md:1
curl -fsSL https://cli.inference.sh | shLOW Dangerous data flow in command pipeline Data Exfiltrationpipeline
Pipeline downloads data from the network and executes it: `curl -fsSL https://fnm.vercel.app/install | bash`. This is a remote code execution pattern. (Note: found in documentation file - may be instructional rather than executable.)
references/cli.md:2
curl -fsSL https://fnm.vercel.app/install | bashLOW Dangerous data flow in command pipeline Data Exfiltrationpipeline
Pipeline downloads data from the network and executes it: `curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.40.0/install.sh | bash`. This is a remote code execution pattern. (Note: found in documentation file - may be instructional rather than executable.)
references/cli.md:6
curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.40.0/install.sh | bashINFO Skill does not specify a license Policy Violationstatic
Skill manifest does not include a 'license' field. Specifying a license helps users understand usage terms.
SKILL.md
What is in the package
| File | Kind | Size |
|---|---|---|
SKILL.md | doc | 12 KB |
references/cli.md | doc | 2.7 KB |
references/node-app-logic.md | doc | 4.8 KB |
references/node-patterns.md | doc | 5.1 KB |
references/node-secrets-oauth.md | doc | 3.0 KB |
references/node-tracking.md | doc | 2.3 KB |
references/python-app-logic.md | doc | 4.9 KB |
references/python-patterns.md | doc | 6.8 KB |
references/python-secrets-oauth.md | doc | 3.3 KB |
references/python-tracking.md | doc | 3.0 KB |
External hosts referenced
astral.sh cli.inference.sh fnm.vercel.app github.com inference.sh raw.githubusercontent.com
Hostnames that appear in the skill's files. Referencing a host is not a finding; compare against the findings above and the skill's stated purpose.
Environment variables referenced
OPENAI_API_KEY API_KEY GOOGLE_OAUTH_CREDENTIALS GOOGLE_SA_CREDENTIALS HF_TOKEN INFSH_API_KEY
Package contents recorded 2026-10-22 at commit becc25649700. The scan above ran against the version current at scan time; the two can differ if the skill has since been updated.
What we checked
- Static analyzer. Pattern rules over SKILL.md and bundled files: known-bad commands, exfiltration URLs, secrets, obfuscation.
- Bytecode analyzer.
- Pipeline analyzer.
- Correlation analyzer.
- Behavioral analyzer. Checks what the skill asks the agent to do against what its description claims it does.
- Trigger analyzer. Looks for instructions that fire on conditions the user did not ask for.
What we did not check
- Runtime behaviour. The skill was not executed; findings come from reading its files.
- Anything fetched at install or run time, including remote scripts, package dependencies, and URLs the skill tells the agent to open.
- Content the analyzers cannot read: images, binaries, and some non-English text.
- Known scanner bypass classes. See the methodology and limitations page.
Sources
- Repository: https://github.com/magentosh/superpowers
- Raw scan result: magentosh-superpowers-sdk-building-apps-scan.json
- Badge for your README:
- Request a re-scan