dynamic-debugger

This file appears to be a benign integration test harness that contains risky operational practices rather than explicit malicious code. The immediate fragment is syntactically broken/missing definitions (reducing runnable risk), but the intended full script would present a moderate supply-chain and execution risk because it imports project-local code by modifying sys.path, writes config to /tmp, and executes project-supplied shell scripts and the generated test program via subprocess.run. No evidence of direct data exfiltration, network callbacks, hard-coded credentials, or obfuscated malware was found in the fragment. Treat the execution of project-provided scripts and files as a primary attack surface and apply repository integrity checks and safer temporary-file and subprocess handling to mitigate risk.

This SKILL.md describes a legitimate debugging skill whose capabilities align with its stated purpose. The primary security concerns stem from the high privileges required for debugging: full filesystem read/write, starting and attaching debuggers to processes, and access to in-memory data and environment variables. Those capabilities can expose sensitive credentials and enable execution of arbitrary code if the skill is enabled or misconfigured. No explicit malicious code, remote exfiltration endpoints, or download-and-execute pipelines are present in the provided text. Risk posture: moderate. The skill should remain disabled by default (as documented), require explicit operator opt-in, run only in trusted development environments, and include operational safeguards (confirmations, readonly modes, network binding checks, and clear instructions to avoid attaching to production processes containing secrets).

This README describes a legitimate dynamic debugging skill that orchestrates dap-mcp and debugger backends. The documentation properly warns that debugging grants broad access to filesystem, process memory, and environment variables. I found no direct evidence of obfuscation, hard-coded credentials, or intentional malware in the provided fragment. The primary security concern is operational: if dap-mcp or the orchestration scripts are misconfigured, exposed to networks, run in CI with secrets, or replaced by malicious binaries, attackers can use the debugger capabilities to read and exfiltrate sensitive data. Before deploying, review the actual server and script implementations for secure defaults (authentication, TLS, bind-to-localhost), apply least-privilege, and isolate debugging sessions.