Security audit

golang-continuous-integration

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly a CI template package, but its AI review workflow would add persistent automation with broad GitHub permissions and mutable remote installs that should be reviewed before use.

Install only after reviewing or removing the AI review workflow. Pin npx packages, GitHub Actions, Go tooling, and the fetched skills repository to immutable versions; remove id-token: write from review jobs; scope actions: read to only the diagnosis job; restrict AI review triggers to trusted contributors; treat PR files, comments, diffs, and logs as untrusted data; and avoid posting raw workflow log excerpts to PRs. The ordinary CI templates are more coherent, but release, Docker publishing, and auto-merge workflows still require branch protection and careful secret management.

A.I.G

Vulnerability Patterns

Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

Findings (2)

T01 · Skill Instruction Hijacking

Error

Location: assets/claude-code-review.yml:16
Finding: Prompt injection exposure in AI review workflow with repository-wide write and log-read privileges
Content: View full analysis
" additional_permissions: | actions: read claude_args: >- --allowedTools "mcp__github_inline_comment__create_inline_comment,mcp__context7__resolve-library-id,mcp__context7__query-docs,Bash(gh pr comment:*),Bash(gh pr diff:*),Bash(gh pr view:*)" ``` The same privileged action pattern is repeated in the `correctness`, `security`, and `quality-depth` jobs. The prompts also direct the agent to load repository-controlled files: ```yaml Note: the PR branch is already checked out in the current working directory. Check project guidelines: @./CLAUDE.md Check contributing guidelines: @./CONTRIBUTING.md Check project description: @./docs/project-summary.md ``` The workflow exposes a log-reading agent after the review jobs: ```yaml ci-diagnosis: name: Review — CI Failure Diagnosis runs-on: ubuntu-latest ...[truncated 4987 chars]
Remediation: View remediation

T03 · Remote Payload Retrieval and Execution

Warning

Location: assets/claude-code-review.yml:47
Finding: Mutable remote dependencies are downloaded and executed without immutable version pinning
Content: View full analysis
``` GitHub Actions throughout the templates are also referenced through mutable major-version tags, for example: ```yaml - uses: actions/checkout@v6 - uses: actions/setup-go@v6 - uses: anthropics/claude-code-action@v1 ``` ### Technical Analysis The `npx skills` invocation does not specify an exact npm package version. Depending on local cache and npm behavior, `npx` may retrieve and execute the current published version of the CLI. The command then retrieves Skill content from a GitHub repository without specifying an imm ...[truncated 2681 chars]
Remediation: View remediation

SkillSpector

Vulnerability Patterns

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (26)

Hidden Instructions

High

Category: Prompt Injection
Confidence: 90% confidence
Finding: The HTML comment contains hidden operational instructions telling the agent/user how to fetch and copy external skills locally, including running npx and creating a symlink. Hidden instructions are risky because they can influence automated tooling or humans in ways not visible in normal rendered output, and here they specifically encourage pulling external content into the trust boundary.
Content: Scanner excerpt · assets/copilot-review-instructions.md (reported line 1)May include surrounding context.
md
<!-- Prerequisites: The skills CLI (listed in the frontmatter install block) can be used to copy skills locally: npx skills add https://github.com/samber/cc-skills-golang --agent github-copilot --skill '*' -y --copy ln -s .agents .copilot

Credential Access

High

Category: Privilege Escalation
Confidence: 70% confidence
Finding: Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
Content: Scanner excerpt · references/repo-security.md (reported line 40)May include surrounding context.
md
For public/open-source repositories: 1. In **Actions permissions** (link above), set **"Fork pull request workflows from outside collaborators"** to **"Require approval for all outside collaborators"** 2. This prevents untrusted forks from running workflows that consume your Actions minutes or access secrets 3. NEVER use `pull_request_target` with untrusted code — it runs with write access to the base repo ## Secrets and Environments

Autonomous Decision Making

Medium

Category: Excessive Agency
Confidence: 80% confidence
Finding: Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
Content: Scanner excerpt · SKILL.md (reported line 158)May include surrounding context.
md
Key advantages over Dependabot: - **`gomodTidy`**: Automatically runs `go mod tidy` after updates - **Native automerge**: No separate workflow needed - **Better grouping**: More flexible rules for grouping PRs - **Regex managers**: Can update versions in Dockerfiles, Makefiles, etc.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 92% confidence
Finding: The skill instructs users to execute npx skills ... without pinning an explicit package version, which can fetch whatever version is current at execution time. In a security-sensitive setup workflow, this creates a supply-chain risk: a compromised or malicious future release of the skills package could run arbitrary code in the user's environment.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 90% confidence
Finding: The workflow triggers broadly on PR opens/updates and review events, yet the file itself notes the permissions apply repository-wide and recommends narrowing to trusted contributors or protected branches. In combination with external dependency fetching, PR comment write access, workflow log access, and OIDC capability, broad trigger scope makes it easier for untrusted contributions to exercise sensitive automation and increases the chance of abuse or accidental data exposure.
Content: No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The workflow grants id-token: write, which enables OIDC token minting, but nothing in the stated code-review behavior requires cloud federation. Unnecessary OIDC capability increases the blast radius of any compromise in the action, prompts, or fetched dependencies, because an attacker may be able to exchange the token with trusting cloud services.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 92% confidence
Finding: The workflow runs npx skills add without pinning the package version, so each PR-triggered run can pull whatever package version is current at execution time. In a CI code-review workflow that has repository-scoped read access, PR comment write access, workflow log access, and an OIDC token, this creates a real supply-chain risk: a compromised or malicious upstream package version could execute attacker-controlled code inside the runner.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 92% confidence
Finding: This is the same unpinned npx skills add pattern in a separate job. Because multiple jobs independently fetch and execute the tool on PR events, the attack surface is multiplied and a malicious package update could compromise several review paths in the same workflow run.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 92% confidence
Finding: This job also executes npx skills add without version pinning, allowing code from an external package registry and remote repository to change between runs. In the security review job, that is especially dangerous because the fetched code influences security-significant output while running with write/commenting capabilities and access to workflow metadata.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 92% confidence
Finding: The same unpinned package execution appears in the quality-depth job, so an attacker controlling or compromising the upstream package/repository could gain code execution in CI. Repeated use across jobs increases exposure and makes the workflow dependent on mutable third-party content for trusted automation decisions.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 87% confidence
Finding: The CI diagnosis job is instructed to inspect workflow logs and post excerpts back to the PR, but there is no hard safeguard at the action point preventing sensitive log content from being copied into a public discussion. Since workflow logs can contain stack traces, tokens accidentally printed by other steps, environment details, or internal URLs, this creates a realistic confidentiality risk if a failed job logs sensitive data.
Content: No source excerpt is available for this finding.

Rp1

Medium

Category: MCP Rug Pull
Confidence: 94% confidence
Finding: The commented prerequisite uses npx skills add ... without pinning an exact package version. If a user follows these setup instructions, they may execute an unexpected or compromised package version from the registry, creating a supply-chain risk during local environment setup.
Content: No source excerpt is available for this finding.

Skill Enumeration

Medium

Category: Agent Snooping
Confidence: 80% confidence
Finding: Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.
Content: Scanner excerpt · assets/copilot-review-instructions.md (reported line 24)May include surrounding context.
md
Cover each area below. Where a dedicated skill is listed, apply its guidance. - **Code style** — formatting, comment quality, idiomatic Go patterns (`.copilot/skills/golang-code-style/SKILL.md`) - **Naming** — packages, types, variables, functions, constants (`.copilot/skills/golang-naming/SKILL.md`) - **Error handling** — wrapping, sentinel errors, log-and-return, swallowed errors (`.copilot/skills/golang-error-handling/SKILL.md`) - **Concurrency** — goroutine lifecycle, mutex usage, channel patterns, context propagation, data races (`.copilot/skills/golang-concurrency/SKILL.md`)

Skill Enumeration

Medium

Category: Agent Snooping
Confidence: 80% confidence
Finding: Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.
Content: Scanner excerpt · assets/copilot-review-instructions.md (reported line 25)May include surrounding context.
md
Cover each area below. Where a dedicated skill is listed, apply its guidance. - **Code style** — formatting, comment quality, idiomatic Go patterns (`.copilot/skills/golang-code-style/SKILL.md`) - **Naming** — packages, types, variables, functions, constants (`.copilot/skills/golang-naming/SKILL.md`) - **Error handling** — wrapping, sentinel errors, log-and-return, swallowed errors (`.copilot/skills/golang-error-handling/SKILL.md`) - **Concurrency** — goroutine lifecycle, mutex usage, channel patterns, context propagation, data races (`.copilot/skills/golang-concurrency/SKILL.md`) - **Code safety** — nil dereference, map/slice aliasing, integer overflows, uninitialized state (`.copilot/skills/golang-safety/SKILL.md`)

Skill Enumeration

Medium

Category: Agent Snooping
Confidence: 80% confidence
Finding: Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.
Content: Scanner excerpt · assets/copilot-review-instructions.md (reported line 26)May include surrounding context.
md
- **Code style** — formatting, comment quality, idiomatic Go patterns (`.copilot/skills/golang-code-style/SKILL.md`) - **Naming** — packages, types, variables, functions, constants (`.copilot/skills/golang-naming/SKILL.md`) - **Error handling** — wrapping, sentinel errors, log-and-return, swallowed errors (`.copilot/skills/golang-error-handling/SKILL.md`) - **Concurrency** — goroutine lifecycle, mutex usage, channel patterns, context propagation, data races (`.copilot/skills/golang-concurrency/SKILL.md`) - **Code safety** — nil dereference, map/slice aliasing, integer overflows, uninitialized state (`.copilot/skills/golang-safety/SKILL.md`) - **Tests** — coverage of new code, test quality, table-driven tests, use of t.Helper() (`.copilot/skills/golang-testing/SKILL.md`)

Skill Enumeration

Medium

Category: Agent Snooping
Confidence: 80% confidence
Finding: Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.
Content: Scanner excerpt · assets/copilot-review-instructions.md (reported line 27)May include surrounding context.
md
- **Code style** — formatting, comment quality, idiomatic Go patterns (`.copilot/skills/golang-code-style/SKILL.md`) - **Naming** — packages, types, variables, functions, constants (`.copilot/skills/golang-naming/SKILL.md`) - **Error handling** — wrapping, sentinel errors, log-and-return, swallowed errors (`.copilot/skills/golang-error-handling/SKILL.md`) - **Concurrency** — goroutine lifecycle, mutex usage, channel patterns, context propagation, data races (`.copilot/skills/golang-concurrency/SKILL.md`) - **Code safety** — nil dereference, map/slice aliasing, integer overflows, uninitialized state (`.copilot/skills/golang-safety/SKILL.md`) - **Tests** — coverage of new code, test quality, table-driven tests, use of t.Helper() (`.copilot/skills/golang-testing/SKILL.md`) - **Performance** — unnecessary allocations, inefficient data structures, missing bounds (`.copilot/skills/golang-performance/SKILL.md`)

Skill Enumeration

Medium

Category: Agent Snooping
Confidence: 80% confidence
Finding: Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.
Content: Scanner excerpt · assets/copilot-review-instructions.md (reported line 28)May include surrounding context.
md
- **Naming** — packages, types, variables, functions, constants (`.copilot/skills/golang-naming/SKILL.md`) - **Error handling** — wrapping, sentinel errors, log-and-return, swallowed errors (`.copilot/skills/golang-error-handling/SKILL.md`) - **Concurrency** — goroutine lifecycle, mutex usage, channel patterns, context propagation, data races (`.copilot/skills/golang-concurrency/SKILL.md`) - **Code safety** — nil dereference, map/slice aliasing, integer overflows, uninitialized state (`.copilot/skills/golang-safety/SKILL.md`) - **Tests** — coverage of new code, test quality, table-driven tests, use of t.Helper() (`.copilot/skills/golang-testing/SKILL.md`) - **Performance** — unnecessary allocations, inefficient data structures, missing bounds (`.copilot/skills/golang-performance/SKILL.md`) - **Security** — injection, auth, crypto misuse, sensitive data exposure, input validation (`.copilot/skills/golang-security/SKILL.md`)

Skill Enumeration

Medium

Category: Agent Snooping
Confidence: 80% confidence
Finding: Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.
Content: Scanner excerpt · assets/copilot-review-instructions.md (reported line 29)May include surrounding context.
md
- **Error handling** — wrapping, sentinel errors, log-and-return, swallowed errors (`.copilot/skills/golang-error-handling/SKILL.md`) - **Concurrency** — goroutine lifecycle, mutex usage, channel patterns, context propagation, data races (`.copilot/skills/golang-concurrency/SKILL.md`) - **Code safety** — nil dereference, map/slice aliasing, integer overflows, uninitialized state (`.copilot/skills/golang-safety/SKILL.md`) - **Tests** — coverage of new code, test quality, table-driven tests, use of t.Helper() (`.copilot/skills/golang-testing/SKILL.md`) - **Performance** — unnecessary allocations, inefficient data structures, missing bounds (`.copilot/skills/golang-performance/SKILL.md`) - **Security** — injection, auth, crypto misuse, sensitive data exposure, input validation (`.copilot/skills/golang-security/SKILL.md`) - **Dependencies** — new imports, license compatibility, known vulnerabilities (`.copilot/skills/golang-dependency-management/SKILL.md`)

Skill Enumeration

Medium

Category: Agent Snooping
Confidence: 80% confidence
Finding: Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.
Content: Scanner excerpt · assets/copilot-review-instructions.md (reported line 30)May include surrounding context.
md
- **Concurrency** — goroutine lifecycle, mutex usage, channel patterns, context propagation, data races (`.copilot/skills/golang-concurrency/SKILL.md`) - **Code safety** — nil dereference, map/slice aliasing, integer overflows, uninitialized state (`.copilot/skills/golang-safety/SKILL.md`) - **Tests** — coverage of new code, test quality, table-driven tests, use of t.Helper() (`.copilot/skills/golang-testing/SKILL.md`) - **Performance** — unnecessary allocations, inefficient data structures, missing bounds (`.copilot/skills/golang-performance/SKILL.md`) - **Security** — injection, auth, crypto misuse, sensitive data exposure, input validation (`.copilot/skills/golang-security/SKILL.md`) - **Dependencies** — new imports, license compatibility, known vulnerabilities (`.copilot/skills/golang-dependency-management/SKILL.md`) - **Documentation** — exported symbols, package docs, README impact (`.copilot/skills/golang-documentation/SKILL.md`)

Skill Enumeration

Medium

Category: Agent Snooping
Confidence: 80% confidence
Finding: Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.
Content: Scanner excerpt · assets/copilot-review-instructions.md (reported line 31)May include surrounding context.
md
- **Code safety** — nil dereference, map/slice aliasing, integer overflows, uninitialized state (`.copilot/skills/golang-safety/SKILL.md`) - **Tests** — coverage of new code, test quality, table-driven tests, use of t.Helper() (`.copilot/skills/golang-testing/SKILL.md`) - **Performance** — unnecessary allocations, inefficient data structures, missing bounds (`.copilot/skills/golang-performance/SKILL.md`) - **Security** — injection, auth, crypto misuse, sensitive data exposure, input validation (`.copilot/skills/golang-security/SKILL.md`) - **Dependencies** — new imports, license compatibility, known vulnerabilities (`.copilot/skills/golang-dependency-management/SKILL.md`) - **Documentation** — exported symbols, package docs, README impact (`.copilot/skills/golang-documentation/SKILL.md`) - **Observability** — logging, metrics, tracing added for new code paths (`.copilot/skills/golang-observability/SKILL.md`)

Skill Enumeration

Medium

Category: Agent Snooping
Confidence: 80% confidence
Finding: Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.
Content: Scanner excerpt · assets/copilot-review-instructions.md (reported line 32)May include surrounding context.
md
- **Tests** — coverage of new code, test quality, table-driven tests, use of t.Helper() (`.copilot/skills/golang-testing/SKILL.md`) - **Performance** — unnecessary allocations, inefficient data structures, missing bounds (`.copilot/skills/golang-performance/SKILL.md`) - **Security** — injection, auth, crypto misuse, sensitive data exposure, input validation (`.copilot/skills/golang-security/SKILL.md`) - **Dependencies** — new imports, license compatibility, known vulnerabilities (`.copilot/skills/golang-dependency-management/SKILL.md`) - **Documentation** — exported symbols, package docs, README impact (`.copilot/skills/golang-documentation/SKILL.md`) - **Observability** — logging, metrics, tracing added for new code paths (`.copilot/skills/golang-observability/SKILL.md`) - **Modernize code** — outdated patterns replaced with Go 1.21+ idioms (`.copilot/skills/golang-modernize/SKILL.md`)

Skill Enumeration

Medium

Category: Agent Snooping
Confidence: 80% confidence
Finding: Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.
Content: Scanner excerpt · assets/copilot-review-instructions.md (reported line 33)May include surrounding context.
md
- **Performance** — unnecessary allocations, inefficient data structures, missing bounds (`.copilot/skills/golang-performance/SKILL.md`) - **Security** — injection, auth, crypto misuse, sensitive data exposure, input validation (`.copilot/skills/golang-security/SKILL.md`) - **Dependencies** — new imports, license compatibility, known vulnerabilities (`.copilot/skills/golang-dependency-management/SKILL.md`) - **Documentation** — exported symbols, package docs, README impact (`.copilot/skills/golang-documentation/SKILL.md`) - **Observability** — logging, metrics, tracing added for new code paths (`.copilot/skills/golang-observability/SKILL.md`) - **Modernize code** — outdated patterns replaced with Go 1.21+ idioms (`.copilot/skills/golang-modernize/SKILL.md`)

Skill Enumeration

Medium

Category: Agent Snooping
Confidence: 80% confidence
Finding: Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.
Content: Scanner excerpt · assets/copilot-review-instructions.md (reported line 34)May include surrounding context.
md
- **Security** — injection, auth, crypto misuse, sensitive data exposure, input validation (`.copilot/skills/golang-security/SKILL.md`) - **Dependencies** — new imports, license compatibility, known vulnerabilities (`.copilot/skills/golang-dependency-management/SKILL.md`) - **Documentation** — exported symbols, package docs, README impact (`.copilot/skills/golang-documentation/SKILL.md`) - **Observability** — logging, metrics, tracing added for new code paths (`.copilot/skills/golang-observability/SKILL.md`) - **Modernize code** — outdated patterns replaced with Go 1.21+ idioms (`.copilot/skills/golang-modernize/SKILL.md`) ## Review Priority

Skill Enumeration

Medium

Category: Agent Snooping
Confidence: 80% confidence
Finding: Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.
Content: Scanner excerpt · assets/copilot-review-instructions.md (reported line 35)May include surrounding context.
md
- **Dependencies** — new imports, license compatibility, known vulnerabilities (`.copilot/skills/golang-dependency-management/SKILL.md`) - **Documentation** — exported symbols, package docs, README impact (`.copilot/skills/golang-documentation/SKILL.md`) - **Observability** — logging, metrics, tracing added for new code paths (`.copilot/skills/golang-observability/SKILL.md`) - **Modernize code** — outdated patterns replaced with Go 1.21+ idioms (`.copilot/skills/golang-modernize/SKILL.md`) ## Review Priority

Rp1

Medium

Category: MCP Rug Pull
Confidence: 70% confidence
Finding: npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.