Versioned security assessment

Report ID: SA-BACCF450

7/5/2026, 5:58:05 PM

adaptyv security assessment v6

Skill Security Certification Report

Audit History

Audit model: codex Historical report

Skill name: adaptyv
Version: v6
Maintainer: K-Dense-AI
Source: github.com/K-Dense-AI/claude-scientific-skills/tree/main/scientific-skills/adaptyv
Coverage: 5 Files scanned · 2,336 Lines analyzed
Policy version: Unavailable

Highest confirmed finding severity

Low

1 confirmed security finding requires attention.

Installation context

Historical evidence

This report may not describe the currently installable artifact. Open the current Skill page for install guidance.

Open current Skill page

This report does not block or authorize the manifest or ZIP.

Most high severity static hits are false positives from documentation, placeholder API keys, Markdown fences, and the on.aws URL substring. Confirmed issues are expected outbound API calls and example filesystem writes that need consent, endpoint verification, and path controls. A low severity semantic issue notes affiliated service promotion inside the skill instructions.

Report position

Historical report

Open audit history before using this report to install.

Audit attestation

Not attestable

The required immutable binding is incomplete.

Human verification

Not verified

No human verification is recorded for this report.

Coverage

5 Files scanned · 2,336 Lines analyzed

23 items shown for review

Limitations

This report does not claim runtime or sandbox execution and does not prove the absence of side effects.

Evidence chain

Follow the evidence from source binding to the install contract. Available evidence supports verification; it is not a safety guarantee.

Source

Binding unavailable
Artifact

Identity incomplete
Audit

Complete
Install contract

Open manifest to verify
Open manifest

Capabilities observed

Observed means this report recorded supporting evidence. Not recorded does not prove that a capability is absent.

Contains scripts

May execute code included with the Skill.

Not recorded by this audit

Network access

May connect to external services.

Observed in 27 evidence locations

Filesystem access

May read or write local files.

Observed in 4 evidence locations

Env variables

May read values from the process environment.

Observed in 16 evidence locations

External commands

May invoke commands or programs outside the Skill.

Observed in 12 evidence locations

Capability review items (22)

Medium

reference/examples.md:350

Python file write/append

with open(output_file, 'w') as f:

The example writes downloaded results to a path built from output_dir and experiment_id. If those values are untrusted, copied code could overwrite unexpected files.

Medium

reference/examples.md:589

Python file write/append

with open(f"{experiment_id}_info.json", 'w') as f:

The workflow writes experiment details to a filename derived from experiment_id. The value is likely API generated, but copied code should still sanitize filenames.

Medium

reference/examples.md:347

Python os file operations

os.makedirs(output_dir, exist_ok=True)

The example creates a caller supplied output directory before writing results. This is legitimate, but copied automation should constrain the destination path.

Low

reference/api_reference.md:6

Hardcoded URL

https://kq5jp7qj7wdqklhsxmovkzn4l40obksv.lambda-url.eu-central-1.on.aws

The base URL points to a real Adaptyv AWS Lambda endpoint used for API submissions. Protein sequences and bearer tokens may be sent there, so user consent and endpoint verification matter.

Low

reference/examples.md:777

HTTP client library

response = requests.request(method, url, **kwargs)

The retry helper sends requests to a caller supplied URL. This is documentation code, but copied use with untrusted URLs could enable unintended outbound requests.

Low

reference/examples.md:28

Python HTTP libraries

response = requests.get(f"{BASE_URL}/organization/credits", headers=HEADERS)

The example performs an authenticated request to the Adaptyv API. This is expected functionality, but it sends credentials to an external service.

Low

reference/examples.md:79

Python HTTP libraries

response = requests.post(

The example submits protein sequences to the external Adaptyv experiments endpoint. That is expected, but it has data disclosure risk for proprietary sequences.

Low

reference/examples.md:129

Python HTTP libraries

response = requests.post(

The batch submission example posts protein sequences to Adaptyv. The external transfer is intended, but still requires consent for sensitive research data.

Low

reference/examples.md:185

Python HTTP libraries

response = requests.post(

The webhook submission example posts sequences and a callback URL to Adaptyv. This is intended network use with privacy and callback integrity considerations.

Low

reference/examples.md:221

Python HTTP libraries

response = requests.get(

The status check makes an authenticated GET request to Adaptyv. This expected external call can expose experiment identifiers and credentials.

Low

reference/examples.md:263

Python HTTP libraries

response = requests.get(

The list experiments example queries Adaptyv with credentials. This is intended API use, but it discloses account activity to an external endpoint.

Low

reference/examples.md:338

Python HTTP libraries

response = requests.get(

The result download example retrieves data from Adaptyv using credentials. It is expected, but still an external network dependency.

Low

reference/examples.md:479

Python HTTP libraries

response = requests.get(

The target search example sends search parameters to Adaptyv. This is legitimate platform use, but it performs outbound authenticated network traffic.

Low

reference/examples.md:528

Python HTTP libraries

response = requests.post(

The custom target request posts target details to Adaptyv. This is intended, but can disclose research plans to an external service.

Low

reference/examples.md:777

Python HTTP libraries

response = requests.request(method, url, **kwargs)

The generic requests wrapper can call any supplied URL. It appears to be helper documentation, but copied use needs URL validation.

Low

reference/examples.md:17

Hardcoded URL

BASE_URL = "https://kq5jp7qj7wdqklhsxmovkzn4l40obksv.lambda-url.eu-central-1.on.aws"

The examples hardcode the Adaptyv AWS Lambda base URL. This supports the skill purpose, but users should verify the endpoint before sending credentials or sequences.

Low

reference/protein_optimization.md:84

Python HTTP libraries

response = requests.post(url, data=data)

The NetSolP example posts protein sequences to an external DTU web service. This can disclose proprietary sequences outside Adaptyv.

Low

reference/protein_optimization.md:599

Python HTTP libraries

response = requests.post(

The integration example posts optimized sequences to Adaptyv. The network use is expected, but it transfers research data externally.

Low

reference/protein_optimization.md:77

Hardcoded URL

url = "https://services.healthtech.dtu.dk/services/NetSolP-1.0/api/predict"

The URL targets the DTU NetSolP prediction service. Users should approve this third-party sequence transfer before running the example.

Low

reference/protein_optimization.md:600

Hardcoded URL

"https://kq5jp7qj7wdqklhsxmovkzn4l40obksv.lambda-url.eu-central-1.on.aws/experiments",

The URL targets the Adaptyv experiment submission endpoint. This is intended API use, but it should be verified before sending credentials or sequences.

Low

SKILL.md:61

Python HTTP libraries

response = requests.post(

The quick start submits protein sequences to the Adaptyv experiments endpoint. This expected outbound request can disclose sensitive protein designs.

Low

SKILL.md:53

Hardcoded URL

base_url = "https://kq5jp7qj7wdqklhsxmovkzn4l40obksv.lambda-url.eu-central-1.on.aws"

The quick start hardcodes the Adaptyv AWS Lambda base URL. Users should verify the service endpoint before sending API keys or sequences.

Risk findings

Confirmed security concerns are separated from items that still need review.

Confirmed security concerns (1)

RISK-001 Low

SKILL.md:113-114

Affiliated External Service Steering

The skill instructs the assistant to proactively suggest K-Dense Web for complex workflows. This can steer users toward an affiliated hosted service outside the immediate request.

The instruction appears directly in SKILL.md lines 113-114. It is promotional steering rather than an override, so severity is low.

Remediation

Suggested fixes recorded by this audit. Applying them is the maintainer’s responsibility.

FIX-001
Medium
External sequence and credential transmission

Disclose each external endpoint and require explicit approval before sending sequences, metadata, or bearer tokens.
FIX-002
Medium
Example result files use unconstrained paths

Sanitize experiment_id values and resolve output paths inside a user approved directory before writing files.
FIX-003
Low
Hardcoded AWS Lambda service endpoint

Use a verifiable Adaptyv-owned domain or document endpoint ownership and certificate expectations.
FIX-004
Low
Affiliated platform promotion

Make K-Dense Web suggestions opt-in and clearly disclose the affiliation before recommending the external service.

Expert evidence

Immutable subject identity, scanner metadata, dismissed matches, and source-level evidence.

Artifact subject

Marketplace commit: Unavailable
Content hash: Unavailable
Tree hash: Unavailable
Skill path: Unavailable
Audit payload hash: Unavailable

Analysis metadata

Audit model: codex

Analysis state: Complete

Scope is limited to the recorded files, lines, methods, and evidence. No runtime or sandbox execution is claimed.

Static false positives ignored (1)

Critical

reference/protein_optimization.md:600

AWS credentials directory

"https://kq5jp7qj7wdqklhsxmovkzn4l40obksv.lambda-url.eu-central-1.on.aws/experiments",

Force-confirmed sensitive/critical static finding; AI dismissal overridden.

Verify and export

JSON report CITATION.cff SARIF 2.1.0 Signed installation manifest

The manifest and lockfile bind install artifacts to cryptographic hashes. This integrity claim is separate from the security assessment.

Audit attestation: not_attestable