Security audit

Debug Checklist - Evomap Asset

Security checks for vulnerabilities and agentic risk

Overview

This is a small C/C++ debugging checklist skill with no hidden persistence, network access, credential use, or privileged behavior, though its bundled header has code-quality risks users should treat cautiously.

Install only if you want a lightweight C/C++ checklist or learning aid. Do not rely on its printed pass/fail messages as a security guarantee, and review or harden checklist.h before compiling it into production or attacker-facing code.

A.I.G

Vulnerability Patterns

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
Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally

Findings (2)

T09 · Insecure Skill Coding Practices

Warning

Location: checklist.h:36
Finding: Unchecked String Pointers Cause Undefined Behavior
Content: View full analysis
Remediation: View remediation

T09 · Insecure Skill Coding Practices

Note

Location: checklist.h:72
Finding: Incomplete Range Validation Produces False Safety Results
Content: View full analysis
array_size) { printf("[warning] Loop may exceed the array boundary\n"); printf("[remediation] Ensure loop_end <= array_size\n"); } else { printf("[pass]\n"); } } ``` The displayed message text is translated into English, while the validation condition and control flow reproduce the audited implementation. ### Technical Analysis The function reports that a range passes whenever `loop_end` is not greater than `array_size`. That single comparison is insufficient to establish that the half-open range `[loop_start, loop_end)` is valid. It does not reject: - A negative `loop_start`, such as `(-1, 10, 10)`. - A negative `loop_end` or `array_size`. - A start position greater than the end position. - A start position greater than the array size. - Integer values whose signed representation is incompatible with the caller's actual array-size type. Consequently, invalid ranges can receive a positive result. For example, both `check_off_by_one(-1, 10, 10)` and `check_off_by_one(20, 10, 10)` print the passing branch even though they do not describe safe array traversals. The helper does not itself access an array, so it does not directly perform an out-of-bounds read or write. The risk arises because its result can mislead code reviewers or developers into treating unsafe indexing logic as valid. ### Attack Path 1. A developer uses `check_off_by_one` to assess a loop range before or during debugging. 2. An invalid range is supplied where `loop_end <= array_size`, but another necessary invariant is false. 3. The helper ...[truncated 999 chars]
Remediation: View remediation
loop_end) { printf("Invalid range: start exceeds end.\n"); return false; } if (loop_end > array_size) { printf("Invalid range: end exceeds array size.\n"); return false; } printf("Range is within the supplied array bounds.\n"); return true; } ``` Additional hardening should include: - Prefer `size_t` for array sizes and indexes when negative values are never valid. - If signed inputs must remain supported, explicitly reject values below zero before converting them to `size_t`. - Return a boolean or structured result rather than only printing a message, allowing callers to enforce failure handling. - Clearly document that validating numeric bounds does not prove that an underlying pointer references an allocation of the declared size. - Add tests for empty ranges, zero-sized arrays, negative signed inputs, reversed ranges, starts beyond the size, ends beyond the size, and maximum representable values. ]]>

SkillSpector

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
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

Findings (2)

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: This header's user-facing strings and comments are entirely in Chinese, including the title at L002 and runtime output beginning at L145. Because the file provides no language choice, opt-in, or documented region-specific constraint, it imposes a specific language/locale in a way that matches the policy's locale-constraint concern.
Content: No source excerpt is available for this finding.

Vague Triggers

Low

Category: Not specified by scanner
Confidence: 82% confidence
Finding: This markdown file describes the skill's purpose and use cases, but it does not define any explicit invocation phrases, constraints, or negative examples. In a skill catalog, such broad descriptions can overlap with many general debugging requests and lead to unintended activation.
Content: No source excerpt is available for this finding.