Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/Liberty91LTD/cti-skillsnpx agentmods add skills/liberty91ltd/cti-skills/mitre-attackWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/liberty91ltd/cti-skills/mitre-attack)<a href="https://agentmods.dev/skills/liberty91ltd/cti-skills/mitre-attack"><img src="https://agentmods.dev/badge/skills/liberty91ltd/cti-skills/mitre-attack/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/liberty91ltd/cti-skills/mitre-attack"><img src="https://agentmods.dev/badge/skills/liberty91ltd/cti-skills/mitre-attack.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00032 | $0.01047 |
| Opus 5 | $0.00016 | $0.00524 |
| Sonnet 5 | $0.00006 | $0.00209 |
| Haiku 4.5 | $0.00003 | $0.00105 |
Grade A, and why
mitre-attack scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
curl -fsSL https://raw.githubusercontent.com/mitre/cti/master/enterprise-attack/enterprise-attack.json \ How it starts
The opening of the file, as written. The whole thing — 111 lines — stays where its author put it; the contents beside it link to each section on GitHub.
MITRE ATT&CK Local Reference
Dataset
Local file: mitre-attack/enterprise-attack.json (~45 MB STIX 2.1 bundle).
First use — if the file is missing, run the bundled download script before any query:
./scripts/download-mitre.sh
The script is idempotent (skips if the file is already present) and supports --force for refresh. If scripts/download-mitre.sh is not present in the install (e.g. for plugin-only deployments), fall back to:
mkdir -p mitre-attack
curl -fsSL https://raw.githubusercontent.com/mitre/cti/master/enterprise-attack/enterprise-attack.json \
-o mitre-attack/enterprise-attack.json
This yields a STIX 2.1 bundle containing all Enterprise ATT&CK objects.
Querying the Dataset
The dataset is a JSON file with a objects array. Each object has a type field.
Find a technique by ID
cat mitre-attack/enterprise-attack.json | python3 -c "
import json, sys
data = json.load(sys.stdin)
tid = 'T1566' # Change as needed
for obj in data['objects']:
refs = obj.get('external_references', [])
for ref in refs:
if ref.get('external_id') == tid:
print(json.dumps(obj, indent=2))
break
"
List all techniques for a tactic
Tactics are mapped via kill_chain_phases[].phase_name:
reconnaissance,resource-development,initial-access,execution,persistenceprivilege-escalation,defense-evasion,credential-access,discoverylateral-movement,collection,command-and-control,exfiltration,impact
Find a threat group
# Groups have type "intrusion-set"
cat mitre-attack/enterprise-attack.json | python3 -c "
import json, sys
data = json.load(sys.stdin)
name = 'APT28' # Change as needed
for obj in data['objects']:
if obj.get('type') == 'intrusion-set':
aliases = obj.get('aliases', [])
if name in aliases or obj.get('name') == name:
print(json.dumps(obj, indent=2))
"
Map group to techniques
Groups link to techniques via relationship objects with relationship_type: "uses".
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 10d ago First seen · 111 lines · 32 tokens per session scan A bc5ae160130c
mitre-attack is a skill published in the GitHub repository Liberty91LTD/cti-skills (17 stars, last pushed 1mo ago), licensed MIT. It adds 32 tokens to every session and 1,047 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
ai-ab-testing
An AI-assisted guide for designing A/B tests, which compare two versions to see which performs better. The description gives no further details about its operation.
ai-accessibility
A Chinese-language assistant for diagnosing accessibility and ease-of-use problems in a page or service. Accessibility means making digital products usable by people with different abilities and needs.
ai-ad-copy-compliance-review
A risk-review assistant for advertising and sales copy. It examines provided material for possible concerns and suggests actions.
ai-agent-browser
A business-diagnosis assistant for marketing, operations, product, and sales questions. It breaks down a goal and produces a summary, findings, action suggestions, and reusable deliverables.
ai-amazon-brand-analytics
An Amazon Brand Analytics assistant for working with Amazon brand-analysis tasks. The description does not provide enough detail about its exact data or outputs.
ai-amazon-dayparting-strategy
An Amazon advertising dayparting assistant. Dayparting means choosing which times of day your ads should run.