The skill is a React Native and Expo troubleshooting guide with no evidence of prompt injection, covert data collection, or unauthorized network activity. Several documented commands can delete caches or simulator data, alter builds, install packages, or terminate processes; users should review them before execution.
1
Files scanned
223
Lines analyzed
14
Review items
0
False positives ignored
Capability review items (11)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
This command block directs the agent to run cache deletion and dependency reinstallation commands. These commands change the project and local development environment.
The cited block contains a kill command using shell command substitution, which can forcibly terminate a local process. The pattern overlaps the dedicated command-substitution finding.
The instruction recursively deletes wildcard-matched directories under /tmp. A broad match can remove Metro or Haste temporary state used by other local work.
The instruction recursively deletes wildcard-matched directories under /tmp. A broad match can remove Metro or Haste temporary state used by other local work.
Most static findings are Markdown fences, documentation links, comments, or standard local debugging guidance rather than executable behavior. Three commands present real local-destructive risk: two broad temporary-directory deletions and a forceful process termination command. No prompt injection, secret collection, data exfiltration, or unauthorized network activity was found.
1
Files scanned
223
Lines analyzed
6
Review items
0
False positives ignored
Capability review items (3)
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 documented command forcefully kills every process listening on port 8081. It can terminate an unrelated local process because it does not verify the process identity.
The command recursively deletes every /tmp entry matching metro-*. Although intended for cache cleanup, it can remove unrelated matching files without confirmation.
The command recursively deletes every /tmp entry matching haste-*. Although intended for cache cleanup, it can remove unrelated matching files without confirmation.
Most static findings are Markdown fences, documentation links, comments, or standard local debugging guidance rather than executable behavior. Three commands present real local-destructive risk: two broad temporary-directory deletions and a forceful process termination command. No prompt injection, secret collection, data exfiltration, or unauthorized network activity was found.
1
Files scanned
223
Lines analyzed
6
Review items
0
False positives ignored
Capability review items (3)
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 documented command forcefully kills every process listening on port 8081. It can terminate an unrelated local process because it does not verify the process identity.
The command recursively deletes every /tmp entry matching metro-*. Although intended for cache cleanup, it can remove unrelated matching files without confirmation.
The command recursively deletes every /tmp entry matching haste-*. Although intended for cache cleanup, it can remove unrelated matching files without confirmation.
Most static findings are Markdown fences, documentation links, comments, or standard local debugging guidance rather than executable behavior. Three commands present real local-destructive risk: two broad temporary-directory deletions and a forceful process termination command. No prompt injection, secret collection, data exfiltration, or unauthorized network activity was found.
1
Files scanned
223
Lines analyzed
6
Review items
0
False positives ignored
Capability review items (3)
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 documented command forcefully kills every process listening on port 8081. It can terminate an unrelated local process because it does not verify the process identity.
The command recursively deletes every /tmp entry matching metro-*. Although intended for cache cleanup, it can remove unrelated matching files without confirmation.
The command recursively deletes every /tmp entry matching haste-*. Although intended for cache cleanup, it can remove unrelated matching files without confirmation.
Most static findings are Markdown fences, official resource links, or local diagnostic steps, not hidden execution or exfiltration. I confirmed the cache deletion commands and force-kill command substitution as medium operational risks because the skill allows Bash. No prompt injection or malicious intent evidence was found.
1
Files scanned
223
Lines analyzed
6
Review items
0
False positives ignored
Capability review items (3)
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 command uses shell command substitution to pass any PID listening on port 8081 to kill -9. In a Bash-enabled skill this can terminate an unrelated local process if run without confirmation.
The command deletes matching directories under /tmp with rm -rf as part of Metro cache cleanup. This is legitimate debugging guidance, but it can remove local files if run automatically or with an unexpected wildcard match.
The command deletes matching directories under /tmp with rm -rf as part of Haste cache cleanup. This is legitimate debugging guidance, but it can remove local files if run automatically or with an unexpected wildcard match.
Most static findings are Markdown fences, official resource links, or local diagnostic steps, not hidden execution or exfiltration. I confirmed the cache deletion commands and force-kill command substitution as medium operational risks because the skill allows Bash. No prompt injection or malicious intent evidence was found.
1
Files scanned
223
Lines analyzed
6
Review items
0
False positives ignored
Capability review items (3)
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 command uses shell command substitution to pass any PID listening on port 8081 to kill -9. In a Bash-enabled skill this can terminate an unrelated local process if run without confirmation.
The command deletes matching directories under /tmp with rm -rf as part of Metro cache cleanup. This is legitimate debugging guidance, but it can remove local files if run automatically or with an unexpected wildcard match.
The command deletes matching directories under /tmp with rm -rf as part of Haste cache cleanup. This is legitimate debugging guidance, but it can remove local files if run automatically or with an unexpected wildcard match.
The skill is a practical React Native and Expo debugging guide with no evidence of prompt injection, credential theft, or malicious network behavior. Several documented commands are legitimate but risky because they delete caches, reset native tooling, inspect logs, or force-kill processes.
The command runs Gradle clean inside the Android project, which is a normal build cleanup. It invokes project build tooling rather than reconnaissance, so the residual risk is low.
The commands read React Native Android logs and adb logcat output. This is legitimate debugging, but device logs may include sensitive application or device information.
Capability review items (12)
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 bash block includes cache clearing, watchman cleanup, rm -rf node_modules, and npm install. These are legitimate React Native recovery steps but can delete local dependencies and execute package lifecycle scripts if run blindly.
This fence closes the build-error command block that includes npx and npm diagnostics. The marker is harmless, but the associated block documents external command execution.
The command kill -9 $(lsof -ti:8081) substitutes process IDs into a force-kill command. It is a common debugging shortcut but can terminate the wrong process if the port ownership is unexpected.
The cited bash block contains command substitution and force-kill process management. It is legitimate but risky enough to require user confirmation before execution.
The command rm -rf /tmp/metro-* deletes temporary Metro cache directories with a wildcard. It is normal cache cleanup but destructive if the glob matches unexpected files.
The command rm -rf /tmp/haste-* deletes Haste cache directories with a wildcard. It is normal React Native cleanup but still destructive filesystem access.
The skill is a legitimate React Native and Expo debugging guide, not a prompt injection or malware attempt. Several static findings are false positives from Markdown fences and documentation links, but the guide does include destructive local reset and process-kill commands that require user confirmation and care.
1
Files scanned
223
Lines analyzed
14
Review items
0
False positives ignored
Capability review items (11)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
This block recommends cache resets including removing node_modules and reinstalling packages. The commands are legitimate debugging steps but can delete local dependency state and run install scripts.
kill -9 $(lsof -ti:8081) uses shell command substitution to terminate any process bound to port 8081. It is a real local process-kill risk even though the target port is fixed.
The surrounding command block includes shell substitution for killing a Metro port process. The syntax is intentional but still a real process termination risk.
rm -rf /tmp/metro-* deletes matching temporary Metro cache directories. The scope is specific, but recursive deletion in /tmp is still a real filesystem risk.
Most static backtick findings are Markdown fences or TypeScript template literals, not executable backtick execution. Real risk remains from Bash-enabled troubleshooting commands that delete caches, reset simulators, inspect device logs, and force-kill Metro processes. No prompt injection, credential exfiltration, or malicious network behavior was found.
The skill presents deletion, simulator reset, emulator stop, and force-kill commands without requiring confirmation before execution.
The cited lines contain concrete destructive commands and no nearby instruction to ask before running them. The intent is troubleshooting, so the severity is medium rather than high.
The commands collect Android device logs through React Native and adb. This is legitimate debugging, but it can expose local device and application data.
The Bash block contains cache-clearing commands that remove node_modules and temporary Metro data. These are legitimate troubleshooting steps, but they can delete local project state if run automatically.
The command substitutes lsof output directly into kill -9 for port 8081. If run blindly, it can terminate any process bound to that port without user confirmation.
The line range contains shell command substitution used to select a PID for force termination. The risk is local process disruption, not template literal execution.
The command recursively deletes temporary Haste cache paths. It should require confirmation because a globbed rm -rf can remove matching local data.
Static false positives ignored (2)
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.
Most static backtick findings are Markdown fences or TypeScript template literals, not executable backtick execution. Real risk remains from Bash-enabled troubleshooting commands that delete caches, reset simulators, inspect device logs, and force-kill Metro processes. No prompt injection, credential exfiltration, or malicious network behavior was found.
The skill presents deletion, simulator reset, emulator stop, and force-kill commands without requiring confirmation before execution.
The cited lines contain concrete destructive commands and no nearby instruction to ask before running them. The intent is troubleshooting, so the severity is medium rather than high.
The commands collect Android device logs through React Native and adb. This is legitimate debugging, but it can expose local device and application data.
The Bash block contains cache-clearing commands that remove node_modules and temporary Metro data. These are legitimate troubleshooting steps, but they can delete local project state if run automatically.
The command substitutes lsof output directly into kill -9 for port 8081. If run blindly, it can terminate any process bound to that port without user confirmation.
The line range contains shell command substitution used to select a PID for force termination. The risk is local process disruption, not template literal execution.
The command recursively deletes temporary Haste cache paths. It should require confirmation because a globbed rm -rf can remove matching local data.
Static false positives ignored (2)
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.
The static analyzer found many shell-command patterns because this skill is a React Native debugging guide with Bash examples. No prompt injection, credential exfiltration, malicious network access, or executable bundled script was found, but several examples delete caches, reset devices, install tools, or kill processes.
1
Files scanned
223
Lines analyzed
6
Review items
3
False positives ignored
Capability review items (3)
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 guide includes commands that remove dependencies and temporary Metro or Haste caches, reset simulators, wipe emulator state, and kill a process on port 8081. These are legitimate mobile debugging steps, but they can delete local project state or interrupt a developer process if run without confirmation.
The cited lines directly show destructive or forceful local operations. The surrounding context is normal debugging guidance, so the concern is operational risk rather than malicious intent.
The skill declares Bash access and provides many local diagnostic commands for package managers, native build tools, device tools, and log collection. This is expected for React Native debugging, but marketplace users should understand that the skill may ask to run commands that affect the local project and connected devices.
The allowed tool list and command blocks are explicit. The commands are common debugging commands, but their breadth creates a real permission and side-effect concern.
System Reconnaissance Alerts Are Debugging Commands
The device and version inspection commands list logs, connected devices, and package versions. These commands support mobile debugging and do not show covert collection, persistence, or exfiltration behavior.
The commands are explicitly scoped to React Native, Expo, iOS simulator, Android emulator, and dependency checks. No evidence shows collection beyond local troubleshooting context.
Static false positives ignored (3)
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.
The static analyzer reported recursive deletion on root or home, but the cited lines target temporary Metro and Haste cache paths under /tmp. This is not evidence of root or home directory deletion, although wildcard deletion still deserves care.
The exact cited commands use /tmp cache prefixes, not root or home paths. The finding is therefore a scanner overmatch.
The static analyzer reported a weak cryptographic algorithm at the description line, but that line only describes crash analysis for React Native debugging. No cryptographic function or hash algorithm is present at that location.
The line is plain metadata text and contains no crypto API or algorithm call. This is a clear false positive.
The hardcoded URLs point to public documentation for React Native, Expo, and Flipper. No code sends data to these URLs, and no evidence of data exfiltration was found.
The URLs appear only in a Resources section. They are user-facing documentation links, not network request code.
This is a documentation-only skill containing debugging guidance for React Native apps. The static analyzer flagged 56 patterns in markdown examples (bash commands, URLs, temp directory paths) that it mistook for executable code. All findings are FALSE POSITIVES. The skill provides no executable code and makes no network requests, file modifications, or system changes. Commands shown are standard React Native/Expo debugging procedures publicly documented in official resources.
This is a documentation-only skill containing debugging guidance for React Native apps. The static analyzer flagged 56 patterns in markdown examples (bash commands, URLs, temp directory paths) that it mistook for executable code. All findings are FALSE POSITIVES. The skill provides no executable code and makes no network requests, file modifications, or system changes. Commands shown are standard React Native/Expo debugging procedures publicly documented in official resources.
This is a pure documentation/markdown skill containing debugging guidance for React Native apps. No executable code, scripts, or network operations. The bash commands shown are standard developer commands publicly documented in React Native/Expo resources.
1
Files scanned
223
Lines analyzed
0
Review items
0
False positives ignored
No confirmed security findings were recorded for this completed audit.
This is a pure documentation/markdown skill containing debugging guidance for React Native apps. No executable code, scripts, or network operations. The bash commands shown are standard developer commands publicly documented in React Native/Expo resources.
1
Files scanned
223
Lines analyzed
0
Review items
0
False positives ignored
No confirmed security findings were recorded for this completed audit.
This is a pure documentation/markdown skill containing debugging guidance for React Native apps. No executable code, scripts, or network operations. The bash commands shown are standard developer commands publicly documented in React Native/Expo resources.
1
Files scanned
223
Lines analyzed
0
Review items
0
False positives ignored
No confirmed security findings were recorded for this completed audit.