Back to skill

Security audit

GitHub Smart Code Digest

Security checks for vulnerabilities and agentic risk

Overview

The skill has a coherent code-review reporting purpose, but it needs review because it uses credentials, shell commands, scheduled publishing, and Feishu Wiki writes in ways that are not tightly scoped.

Review this skill before installing. Use only trusted, pinned dependency versions; restrict GitHub and Feishu credentials to the specific repositories and Wiki space; avoid running it on sensitive repositories until shell input validation and private per-run temp files are added; and confirm who can read the target Feishu Wiki pages before enabling cron publishing.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
Findings (3)

T09 · Insecure Skill Coding Practices

Error
Location
workflow.json:58
Finding

Shell Command Injection Through Unvalidated Workflow Parameters

Content
View full analysis
=${since}\" --json number,title,author,createdAt,changedFiles,additions,deletions --jq '.[]'", "get_diff": "gh pr diff ${pr_number} --repo ${repo} > /tmp/pr_${pr_number}.diff", "get_commits": "gh api repos/${repo}/commits?since=${since_unix}&until=${until_unix} --jq '.[].commit.message'" } ``` ### Technical Analysis The workflow constructs shell command strings by directly interpolating `${repo}`, `${since}`, `${pr_number}`, `${since_unix}`, and `${until_unix}`. The project does not define validation, allowlisting, or shell-safe escaping for these values. Several values originate from invocation arguments or user-controlled configuration. If the workflow engine expands these templates before passing the resulting text to a shell, shell metacharacters, redirections, separators, or command substitutions in a crafted value will be interpreted as shell syntax. The double quotes around `${since}` are insufficient protection because command substitution remains active inside double-quoted shell strings. The unquoted `${repo}` and `${pr_number}` substitutions are additionally exposed to word splitting, glob expansion, and shell control operators. ### Attack Path 1. An attacker gains the ability to supply a repository, date, or related workflow parameter through CLI arguments, configuration, or an automation system invoking the Skill. 2. The attacker supplies a value containing shell syntax instead of the expected repository, integer, or date format. 3. The workflow engine interpolates the value into one of the shell command templates. 4. The generated command is executed by a shell. 5. The shell interprets the injected syntax and executes attacker-selected commands with the privileges and environment of ...[truncated 988 chars]
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
workflow.json:54
Finding

Predictable Shared Temporary Files Allow Symlink Attacks and Cross-Run Data Injection

Content
View full analysis
=${since}\" --json number,title,author,createdAt,changedFiles,additions,deletions --jq '.[]'", "get_diff": "gh pr diff ${pr_number} --repo ${repo} > /tmp/pr_${pr_number}.diff", "get_commits": "gh api repos/${repo}/commits?since=${since_unix}&until=${until_unix} --jq '.[].commit.message'" } }, { "step": 2, "skill": "code-review-skill", "action": "batch_review_prs", "description": "对每个 PR 执行 5 维度 AI 代码审查,过滤低置信度问题", "inputs": { "pr_list": "/tmp/digest_prs.json", "diff_files": "/tmp/pr_*.diff", "min_confidence": 80 }, "outputs": { "review_results": "/tmp/digest_reviews.json", "summary_stats": "/tmp/digest_stats.json" } ``` The Chinese `description` is source content reproduced verbatim; it does not alter the technical finding. ### Technical Analysis The workflow stores artifacts under fixed, globally predictable names in the shared `/tmp` directory. The diff command uses ordinary shell redirection: ```bash > /tmp/pr_${pr_number}.diff ``` Ordinary shell redirection generally follows an existing symbolic link. A local user able to write to `/tmp` can pre-create a predictable path as a symbolic link to another file. When the workflow runs, the target may be truncated or overwritten if the OpenClaw account has permission to write to it. The input pattern `/tmp/pr_*.diff` also consumes every matching file in the shared directory rather than an explicit list of files created by the current run. This permits stale files, attacker-created files, and files from concurrent runs to enter the review process. Fixed names such as `/tmp/digest_prs.json`, `/tmp/digest_re ...[truncated 1808 chars]
Remediation
View remediation

T08 · Insecure Dependencies

Warning
Location
README.md:47
Finding

External Skill Dependencies Are Installed Without Version or Integrity Pinning

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (6)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The README promotes automatic publication of code review results and repository-derived data to Feishu Wiki, but it does not warn users that PR metadata, file paths, code excerpts, comments, or potentially sensitive findings may be transferred to a third-party documentation platform. In a code-review aggregation skill, this omission increases the risk of unintended internal code, security issue, or developer-activity disclosure through routine automation.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill clearly publishes AI review results, repository metadata, and issue summaries into a Feishu Wiki space, but the description does not warn users that this may expose internal engineering details to everyone with access to that space. In this context, the report can include sensitive code-quality findings, filenames, PR titles, and security-related observations, creating an information disclosure risk if the Wiki audience is broader than intended.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

This manifest is entirely oriented around Chinese output, beginning with a Chinese-only description of the skill's purpose. Because the file does not offer any language choice, opt-in, or justification for a Chinese-only locale, it creates a natural-language locale policy concern under the requirement to avoid forcing a specific language without user choice.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The page content template hardcodes Chinese headers, labels, status text, and footer content for the generated report. Since the skill publishes user-facing content and provides no locale toggle or documented regional limitation, it effectively forces a specific language on users, which matches the policy-violation category.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
76% confidence
Finding

The skill documentation and user-facing examples are entirely in Chinese, including labels and operational guidance, with no indication that users may choose another language or that the skill is intentionally restricted to a Chinese-speaking environment. Under the language/locale policy, forcing a specific language without opt-in can be a natural-language policy issue.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
95% confidence
Finding

The manifest description and trigger phrases are entirely specified in Chinese, and the rest of the skill instructions and examples are also Chinese-only. There is no indication that users may choose another language or that the locale restriction is intentional and documented as region-specific.

Content

No source excerpt is available for this finding.