Crispr Grna Designer

• Outcome: Review
• Latest audit: May 29, 2026 · 9:25 PM UTC
• Version: 0.1.0

Overview

This skill does not look malicious, but it presents CRISPR guide design as real while the included implementation silently uses mock sequence data and simulated off-target checks.

Install only for demonstration, education, or code review unless you independently verify the implementation. Do not use its guide RNA, efficiency, or off-target outputs for experiments until it performs real gene/exon retrieval, real genome-wide off-target analysis, clear mock-mode labeling, tests/benchmarks, and pinned dependency resolution.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration

Findings (13)

Intent-Code Divergence

Medium

• Confidence: 95% confidence
• Finding: The function advertises Ensembl-backed sequence retrieval but always returns hard-coded mock data, which can cause users to design gRNAs against the wrong locus while believing the output is biologically valid. In a genome-editing skill, silent substitution of fake reference data is safety-relevant because it can directly lead to incorrect experimental designs and misleading downstream decisions.

Intent-Code Divergence

Medium

• Confidence: 96% confidence
• Finding: The code claims off-target prediction but only returns a hard-coded simulated hit for one sequence pattern, giving users false assurance that guides have been screened. For CRISPR design, inaccurate off-target analysis is a material safety issue because users may select guides with unrecognized genomic off-target activity.

Vague Triggers

Medium

• Confidence: 81% confidence
• Finding: The trigger description is broad enough that the skill could activate for many genome-editing or analysis requests without clear limits. In a high-risk skill that can invoke scripts, call external APIs, and write files, overbroad activation increases the chance of unintended execution, misuse in unsafe contexts, or being selected when a safer/read-only skill would suffice.

Unpinned Dependencies

Low

• Category: Supply Chain

Content

bio
biopython
dataclasses
numpy

• Confidence: 97% confidence
• Finding: The dependency `bio` is unpinned, so installs can pull different versions over time and unexpectedly introduce breaking changes or newly disclosed vulnerabilities. In a bioinformatics skill that may parse complex scientific data and potentially interact with external resources, non-reproducible dependency resolution increases supply-chain risk.

Unpinned Dependencies

Low

• Category: Supply Chain

Content

bio
biopython
dataclasses
numpy
pandas

• Confidence: 98% confidence
• Finding: `biopython` is listed without a version pin, which makes builds non-reproducible and can silently pull vulnerable or incompatible releases. Because this package is central to sequence-processing workflows, unexpected upgrades could directly affect parsing and downstream security posture.

Unpinned Dependencies

Low

• Category: Supply Chain

Content

bio
biopython
dataclasses
numpy
pandas
pysam

• Confidence: 93% confidence
• Finding: `dataclasses` is unpinned, so dependency resolution is not deterministic and future installs may behave differently. This is mainly a supply-chain hygiene issue rather than an immediately exploitable flaw, but it still weakens build integrity.

Unpinned Dependencies

Low

• Category: Supply Chain

Content

bio
biopython
dataclasses
numpy
pandas
pysam
requests

• Confidence: 98% confidence
• Finding: `numpy` is unpinned, allowing potentially vulnerable or incompatible versions to be installed without notice. Given the package's widespread use and prior security advisories, this increases exposure to supply-chain and dependency-drift issues.

Unpinned Dependencies

Low

• Category: Supply Chain

Content

biopython
dataclasses
numpy
pandas
pysam
requests

• Confidence: 98% confidence
• Finding: `pandas` is unpinned, which permits arbitrary future versions during installation and weakens reproducibility. Since data-processing libraries often handle complex inputs, unexpected upgrades can introduce both security and reliability regressions.

Unpinned Dependencies

Low

• Category: Supply Chain

Content

dataclasses
numpy
pandas
pysam
requests

• Confidence: 97% confidence
• Finding: `pysam` is unpinned, creating non-deterministic builds and increasing the chance of pulling a problematic release. In genomics tooling, where file parsing can be complex and performance-critical, version drift is a meaningful security and stability concern.

Unpinned Dependencies

Low

• Category: Supply Chain

Content

numpy
pandas
pysam
requests

• Confidence: 99% confidence
• Finding: `requests` is unpinned, so installations may fetch versions with different security behavior or known flaws. Because this library is commonly used for outbound network access, dependency drift can materially affect confidentiality and transport security.

Known Vulnerable Dependency: biopython — 1 advisory(ies): CVE-2025-68463 (Biopython is vulnerable to doctype XML external entity (XXE) injection through B)

Low

• Category: Supply Chain
• Confidence: 82% confidence
• Finding: The requirements file includes `biopython` without a version restriction, and the static analysis indicates at least one XXE-related advisory affecting some versions. If the skill processes attacker-controlled biological data or XML-like inputs through vulnerable code paths, this could lead to file disclosure or server-side request behavior.

Known Vulnerable Dependency: numpy — 10 advisory(ies): CVE-2014-1859 (Numpy arbitrary file write via symlink attack); CVE-2021-41495 (NumPy NULL Pointer Dereference); CVE-2021-33430 (NumPy Buffer Overflow (Disputed)) +7 more

Critical

• Category: Supply Chain
• Confidence: 71% confidence
• Finding: `numpy` is included without any version constraint, and the scanner reports multiple historical advisories across different versions. The risk is real at the dependency-management level because an unsafe version could be installed, but exploitability depends heavily on the exact version and whether affected functionality is used.

Known Vulnerable Dependency: requests — 10 advisory(ies): CVE-2014-1830 (Exposure of Sensitive Information to an Unauthorized Actor in Requests); CVE-2024-47081 (Requests vulnerable to .netrc credentials leak via malicious URLs); CVE-2024-35195 (Requests `Session` object does not verify requests after making first request wi) +7 more

High

• Category: Supply Chain
• Confidence: 88% confidence
• Finding: `requests` is present without a version pin, and several advisories affect specific releases, including issues related to credential leakage and TLS/session verification behavior. In a skill that may fetch genome annotations, references, or external datasets, a vulnerable HTTP client could expose secrets or weaken transport integrity when interacting with attacker-controlled URLs.

VirusTotal

61/61 vendors flagged this skill as clean.

Static analysis

No suspicious patterns detected.