Security audit

Code Remediation with Gomboc.ai Community Edition

Security checks for vulnerabilities and agentic risk

Overview

The skill is purpose-built for code remediation, but it recommends automated commit/push and persistent MCP workflows with a token in ways users should review carefully before enabling.

Install only if you are comfortable with Gomboc receiving repository-related scan context and with the workflows you enable. Start with scan-only use, store GOMBOC_PAT in a secret manager or CI secret, avoid putting literal tokens in shell profiles, pin and verify any Docker image, stop the MCP service when finished, and require pull request review before applying, committing, or pushing generated fixes.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (9)

Intent-Code Divergence

Medium

Confidence: 91% confidence
Finding: The documentation asserts the token is read-only and cannot modify configurations, yet the skill also promotes remediation workflows that change local code and can create commits. Even if the remote API token is technically limited, the combined workflow can still produce impactful repository changes, so the security claims are misleading and may cause users to underestimate risk.

Intent-Code Divergence

Medium

Confidence: 95% confidence
Finding: Claiming the skill has 'no destructive capabilities' conflicts with documented commands that can remediate code, commit changes, and in some examples push them. This can mislead operators and agents into running file- and repo-modifying actions without appropriately evaluating the blast radius.

Missing User Warnings

Medium

Confidence: 86% confidence
Finding: The README advertises commands that automatically modify source code and can create commits, but it does not explicitly instruct users to review the generated changes before applying or committing them. In an agentic or CI context, this can normalize blind execution of remediation steps and lead to unintended code changes, unsafe fixes, or mass automated commits being accepted without human validation.

Missing User Warnings

Medium

Confidence: 88% confidence
Finding: The skill documents direct code modification, commit, and push workflows without prominent safety warnings about altering local files or repositories. In an agent context, this increases the chance of unintended autonomous changes, especially if users assume the tool is purely advisory.

Missing User Warnings

Medium

Confidence: 89% confidence
Finding: The workflow documentation explicitly recommends running an auto-remediation command with `--commit --push`, which can modify repository contents and publish changes automatically from CI. In the context of a pull request or push-triggered workflow, this can lead to unreviewed or unexpected code changes being committed, especially if users copy the example without understanding the repository-modifying behavior.

Missing User Warnings

Medium

Confidence: 84% confidence
Finding: The scheduled remediation example encourages unattended recurring execution of remediation workflows but does not warn that such jobs may repeatedly modify code over time without a human in the loop. This increases the risk of surprise changes, workflow churn, or accidental propagation of bad fixes if the remediation logic produces undesirable edits.

Missing User Warnings

Medium

Confidence: 94% confidence
Finding: The guide promotes commands that can automatically modify source code and even create commits or pull requests, but it does not warn users about those side effects or recommend review before applying changes. In an agent-integration context, undocumented write/commit behavior increases the risk of unintended repository changes, unsafe automated remediations, or actions being triggered with more authority than the user expects.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The documentation instructs users to provide a personal access token to the agent integration without any guidance on secure storage, least-privilege scoping, rotation, or avoiding disclosure in logs and configs. In an MCP/agent setting, mishandled credentials can be exposed through environment dumps, debug output, shared config files, or overprivileged token use, leading to unauthorized access.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The setup guide instructs users to place a personal access token directly into shell commands and persist it in their shell profile without any warning about credential sensitivity. This can leak the token through shell history, shared terminals, screen recordings, copied dotfiles, backups, or accidental commits, making unauthorized API access more likely.

Static analysis

No suspicious patterns detected.