terraform-consumer-design

This skill's stated purpose (generate Terraform design artifacts and manage them via GitHub issues) is coherent with its capabilities: reading an issue template, mapping user input, creating GitHub issues, running speckit scripts, and committing artifacts. The primary security concerns come from the scope of filesystem and environment access, use of the gh CLI (which uses tokens), and the transitive trust introduced by invoking multiple speckit subagents and writing gh-issue.json for downstream agents. These create realistic opportunities for credential exposure or accidental leakage of sensitive configuration into Git history or issue comments if outputs are not sanitized. The default 'Fully Autonomous' autonomy setting conflicts with the explicit human approval step and raises potential for actions to be taken without proper review. No explicit malicious code, obfuscated payloads, or external exfiltration endpoints are present in the manifest itself, so confirmed malware is unlikely; however, the skill is medium-risk in operational use and requires careful handling of credentials, explicit sanitization rules, and least-privilege credentials for the gh CLI and any subagents. Recommend: require scoped, short-lived GitHub tokens, avoid echoing environment values into issues or committed files, vet/pin the speckit subagents and any scripts in .specify/scripts/bash, and enforce human approval before any implementation-phase actions.