terraform-tactics

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly instructs reading, extracting, and outputting secrets from Terraform state/credentials (jq/cat/grep/terraform output -raw) and using stolen tokens in API requests, which requires handling secret values that could be exfiltrated verbatim.

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 1.00). These URLs include explicit attacker-controlled endpoints and a malicious Git module source (placeholders and attacker domains used for exfiltration and supplying malicious Terraform modules), which are high-risk indicators for supply‑chain malware and RCE/exfiltration.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This repository is an explicit, actionable offensive guide: it contains templates and commands for credential theft, data exfiltration, remote code execution, backdoors, state tampering, and supply‑chain attacks against Terraform environments.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). 该技能的运行流程要求在本地/远程 State 后端下载后读取并解析 terraform.tfstate（包括 jq 提取），这些 State 内容可由外部攻击者通过远程后端写入或投毒来影响，因此属于对“外人可提交/投毒的自由文本”的读取暴露。

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 1.00). The skill explicitly instructs system-wide discovery (find /), reading credential files, extracting secrets, and modifying Terraform state and configs (including injecting providers, local-exec/external RCE and adding backdoor users), all of which actively modify or enable modification of the machine/infrastructure state.