Most external-command findings are false positives from Markdown fences, inline code, or JavaScript template literals. Confirmed risks include Dice.com network automation, persistent browser state, shared temp-directory logging, and Chromium launch flags that disable sandboxing. The skill also has contextual privacy risk because it can upload resume documents and submit job applications.
The Puppeteer launch arguments include --no-sandbox and --disable-setuid-sandbox. This disables important Chromium sandbox protections while browsing third-party pages.
The skill instructs agents to use Dice Easy Apply flows, upload resume or cover-letter PDFs, and submit applications. This can transmit personal data and perform account actions on a third-party service.
The workflow explicitly covers Easy Apply use, verified resume upload, optional cover-letter upload, and submission. The risk is contextual rather than a simple code pattern.
The skill recommends a persistent Chromium profile so login can be performed once and reused. This may retain cookies and authenticated session state on disk.
The file states that a persistent browser profile is used for manual login reuse and shows the profile path in configuration. Session persistence is the intended behavior.
Capability review items (10)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
The skill stores or reuses a Chromium profile under $HOME/.cache/dice-chrome. A persistent profile can retain session cookies and account state on disk.
The skill stores or reuses a Chromium profile under $HOME/.cache/dice-chrome. A persistent profile can retain session cookies and account state on disk.
The skill directs state and results to /tmp/dice-easyapply-daily. Job history in a shared temp directory can expose workflow data or collide with other users.
The skill directs state and results to /tmp/dice-easyapply-daily. Job history in a shared temp directory can expose workflow data or collide with other users.
The skill directs state and results to /tmp/dice-easyapply-daily. Job history in a shared temp directory can expose workflow data or collide with other users.
The skill directs state and results to /tmp/dice-easyapply-daily. Job history in a shared temp directory can expose workflow data or collide with other users.
The skill directs state and results to /tmp/dice-easyapply-daily. Job history in a shared temp directory can expose workflow data or collide with other users.
The skill hardcodes Dice.com URLs for search and application workflows. This is expected for the skill, but it is real third-party network activity tied to job applications.
The skill hardcodes Dice.com URLs for search and application workflows. This is expected for the skill, but it is real third-party network activity tied to job applications.
The skill hardcodes Dice.com URLs for search and application workflows. This is expected for the skill, but it is real third-party network activity tied to job applications.
Most external-command findings are false positives from Markdown fences, inline code, or JavaScript template literals. Confirmed risks include Dice.com network automation, persistent browser state, shared temp-directory logging, and Chromium launch flags that disable sandboxing. The skill also has contextual privacy risk because it can upload resume documents and submit job applications.
The Puppeteer launch arguments include --no-sandbox and --disable-setuid-sandbox. This disables important Chromium sandbox protections while browsing third-party pages.
The skill instructs agents to use Dice Easy Apply flows, upload resume or cover-letter PDFs, and submit applications. This can transmit personal data and perform account actions on a third-party service.
The workflow explicitly covers Easy Apply use, verified resume upload, optional cover-letter upload, and submission. The risk is contextual rather than a simple code pattern.
The skill recommends a persistent Chromium profile so login can be performed once and reused. This may retain cookies and authenticated session state on disk.
The file states that a persistent browser profile is used for manual login reuse and shows the profile path in configuration. Session persistence is the intended behavior.
Capability review items (10)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
The skill stores or reuses a Chromium profile under $HOME/.cache/dice-chrome. A persistent profile can retain session cookies and account state on disk.
The skill stores or reuses a Chromium profile under $HOME/.cache/dice-chrome. A persistent profile can retain session cookies and account state on disk.
The skill directs state and results to /tmp/dice-easyapply-daily. Job history in a shared temp directory can expose workflow data or collide with other users.
The skill directs state and results to /tmp/dice-easyapply-daily. Job history in a shared temp directory can expose workflow data or collide with other users.
The skill directs state and results to /tmp/dice-easyapply-daily. Job history in a shared temp directory can expose workflow data or collide with other users.
The skill directs state and results to /tmp/dice-easyapply-daily. Job history in a shared temp directory can expose workflow data or collide with other users.
The skill directs state and results to /tmp/dice-easyapply-daily. Job history in a shared temp directory can expose workflow data or collide with other users.
The skill hardcodes Dice.com URLs for search and application workflows. This is expected for the skill, but it is real third-party network activity tied to job applications.
The skill hardcodes Dice.com URLs for search and application workflows. This is expected for the skill, but it is real third-party network activity tied to job applications.
The skill hardcodes Dice.com URLs for search and application workflows. This is expected for the skill, but it is real third-party network activity tied to job applications.
The static analyzer over-reported Markdown backticks as Ruby shell execution, and those specific findings are false positives. The skill still has medium operational risk because it directs agents to browse Dice.com with Puppeteer, reuse a persistent authenticated Chromium profile, upload personal resume files, write state logs, and launch Chromium with sandbox-disabling flags.
The skill instructs agents to use Dice Easy Apply, upload a verified resume and cover letter, continue to review, and submit applications. This is legitimate for the stated purpose, but it can disclose personal information or submit applications without sufficient human review if operators do not keep the guardrails.
The lines explicitly describe Easy Apply submission and document upload. The risk is operational and privacy-related rather than confirmed malicious intent.
Capability review items (4)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
The Puppeteer launch arguments include --no-sandbox and --disable-setuid-sandbox while browsing third-party web pages. These flags reduce Chromium process isolation and increase impact if a visited page exploits the browser process.
The sandbox-disabling flags are directly visible in the launch example. They are common in containers, but still materially reduce browser isolation.
Persistent Profile and State Files Retain Sensitive Context
The workflow recommends a persistent Chromium profile plus state and log files under a temporary directory. These files may retain login session artifacts, application history, job URLs, and document-use metadata if permissions and cleanup are not managed.
The skill explicitly recommends a persistent browser profile and local state/log paths. The content also warns against credential storage, so this is a cautionary risk, not evidence of theft.
The hardcoded URLs point to Dice.com search and job-application pages that match the skill purpose. No evidence found of unrelated endpoints, webhook delivery, encoded payload exfiltration, or covert network destinations.
The domains and paths are consistent with Dice job search and application flows. This supports dismissing the hardcoded URL findings as expected network use.
Environment Variables Used for Local Configuration
The environment variable examples configure document paths, browser binary path, profile path, search limits, and dry-run settings. No evidence found that environment variables are sent to an external service or used to collect secrets.
The environment access appears limited to local configuration for Puppeteer and file selection. The same file explicitly says credentials should not be stored in scripts or logs.
Static false positives ignored (1)
These static matches were dismissed by semantic review or matched schema-only tokens, so they are shown for transparency but do not drive the quality score.
False Positive: Markdown Backticks Reported as Shell Execution
The static analyzer flagged many Markdown code fences and inline code spans as Ruby shell backtick execution. The reviewed file contains documentation examples, URLs, JavaScript, bash snippets, and inline labels; no Ruby backtick execution was found.
The cited locations are Markdown fences or inline code examples, not executable Ruby source. This directly explains the bulk of the external_commands static findings.
This skill automates Dice.com job applications using Puppeteer browser automation. Static analysis flagged 59 potential issues, many of which are false positives from the documentation format. Actual security concerns include environment variable access for file paths, temporary file usage, and platform ToS compliance. The skill is not malicious but presents elevated risk due to automated job application behavior that may trigger account penalties on Dice.com.
1
Files scanned
197
Lines analyzed
10
Review items
1
False positives ignored
Capability review items (6)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
Environment Variable Access for Sensitive File Paths
The skill references environment variables (RESUME_PDF, COVER_PDF, CHROME_PROFILE) that point to sensitive user files. While the skill explicitly warns against storing credentials in scripts, the design requires these environment variables to be set with paths to private documents.
Environment variables pointing to user files (resume, cover letter) are legitimate configuration patterns, but improper handling could expose sensitive documents.
The skill accesses ~/.cache/dice-chrome for persistent browser login state. This is a standard pattern for browser automation requiring manual authentication.
Chrome profile storage in ~/.cache is standard behavior; the skill uses this for credential-free authentication.
Static false positives ignored (1)
These static matches were dismissed by semantic review or matched schema-only tokens, so they are shown for transparency but do not drive the quality score.
Static Analysis False Positives - Documentation Format
The static analyzer flagged 'Ruby/shell backtick execution' at multiple lines in SKILL.md, but this is a Markdown documentation file containing example code snippets, not executable Ruby code. The backtick pattern detection was triggered by code fence formatting.
SKILL.md is a documentation file; no backtick command execution exists. Static scanner misidentified code fences as execution patterns.