Content
57%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The body is a well-organized, factually rich actor profile whose emulation and detection sections include genuinely concrete commands, registry keys, and CVEs. Its weaknesses are the absence of validation checkpoints in a workflow built around risky operations (capping workflow clarity at 3), and a monolithic single-file structure that inlines encyclopedic detail instead of offloading it to reference files.
Suggestions
Add explicit validation checkpoints to the emulation workflow — e.g. re-confirm scope/RoE before credential access or persistence phases, and verify with the blue cell that detections fired (DNS-tunneling alerts, password-filter registry changes) before proceeding.
Split the encyclopedic detail (campaign histories, signature-tooling catalog, full per-tactic TTP mapping, detection guidance) into references/ files (e.g. campaigns.md, tooling.md, detection.md) and keep SKILL.md as a lean overview with clearly signaled one-level-deep links.
Trim the 'Attribution & motivation' and 'Notable campaigns' prose and remove the duplication between the frontmatter mitre_attack list and the body's technique-by-technique enumeration to reduce token cost without losing operational value.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense and largely free of fluff, but it restates the entire ATT&CK G0049 catalog — attribution narrative, multi-paragraph campaign histories, and ~75 techniques already enumerated in the frontmatter mitre_attack list — which is encyclopedic knowledge Claude already has. This matches anchor 3 (mostly efficient but includes content that could be tightened or trimmed), and is not 4 because whole sections ("Attribution & motivation", "Notable campaigns") add background rather than operational value. | 3 / 5 |
Actionability | Concrete, executable specifics appear where they matter: the exact LOLBin recon sequence ("whoami, net user/net group, ipconfig /all, netstat -an, net accounts, sc query"), precise registry paths ("HKLM\SYSTEM\CurrentControlSet\Control\Lsa\Notification Packages"), CVE identifiers, and per-tactic tool mappings. It sits between anchors 3 and 4: guidance is mostly concrete, but emulation steps like "use the phishing/social-engineering skill" and "Configure your C2/Sliver profile" lack paths, commands, or examples — noticeable but minor gaps. | 4 / 5 |
Workflow Clarity | The tactic-ordered emulation guidance and the illustrative chain ("phish → loader → DNS/HTTP C2 → discovery → credential access → escalation → lateral access → staging → mail/DNS transfer") provide a rough sequence, but there are no validation checkpoints (no scope re-verification, no per-phase verification, no detection-confirmation steps). Because the operations are destructive/exploitative (credential dumping, kernel exploits, persistence), the guideline capping workflow clarity at 3 applies; the sequence itself keeps it above anchor 2. | 3 / 5 |
Progressive Disclosure | The file is well-sectioned with clear ## headers, but it is a single ~19KB monolith with no bundle files at all — campaign histories, the signature-tooling catalog, per-tactic TTP detail, and detection guidance are all inlined where reference files would keep the SKILL.md an overview. This matches anchor 3 (some structure, but content that should be separate is inline with no references to signal). | 3 / 5 |
Total | 13 / 20 Passed |