NIST AI RMF: MAP 2.2
Provision nist.MAP-2.2 · AI Risk Management Framework 1.0, subcategories · subcategory
Information about the AI system’s knowledge limits and how system output may be utilized and overseen by humans is documented. Documentation provides sufficient information to assist relevant AI actors when making informed decisions and taking subsequent actions.
The text is as taken from the source. In the source: https://airc.nist.gov/airmf-resources/playbook/map/#Map%202.2 (the subcategory, in the NIST AI RMF Playbook (anchor found in the page); linked 2026-11-10). The instrument: AI Risk Management Framework 1.0, subcategories, a United States government work; freely reusable; read 2026-11-10.
Who cites it
| Indicator | The fact it asks for | Kind of link | The assertion |
|---|---|---|---|
I21 | Its autonomy limits and the tools it may use are defined | part: partly evidenced | Asserted in the mapping by gamma-ray@web-summit.dinis; the written basis is not recorded yet. |
I22 | Its limits and known failure cases are documented for those who use it | same: evidenced | Asserted in the mapping by gamma-ray@web-summit.dinis; the written basis is not recorded yet. |
A link is our assertion, not the instrument's. "Same" carries the evidence to the provision in full. "Part" marks the provision as partly evidenced: the fact is one piece of what it asks. "Negation" points the other way.
The evidence that reaches it
| Card | Through | This provision is | The evidence | How far to trust it |
|---|---|---|---|---|
| RiskMandate Publisher | I21 | partly evidenced | Source | π Moderate · S1·R5·A1 ×2 |
| The PrompTugal.news agent team | I21 | partly evidenced | Source | π Moderate · S1·R5·A1 ×2 |
| AI Reviewer | I22 | evidenced | "The AI Reviewer doesn't review pull requests whose changes" Source | π Moderate · S1·R5·A1 ×3 |
| AISLE platform | I21 | partly evidenced | "You decide how far AISLE goes on its own" Source | π Moderate · S1·R5·A1 ×3 |
| AISLE platform | I22 | evidenced | "When findings can't be safely patched, it provides clear remediation advice instead." Source | π Moderate · S1·R5·A1 ×3 |
| ByTheLaw | I21 | partly evidenced | "não pratica, nem pratica em nome do Utilizador, quaisquer atos processuais" Source | π Moderate · S1·R5·A1 ×3 |
| ByTheLaw | I22 | evidenced | "Files attached inside the email are not opened" Source | π Moderate · S1·R5·A1 ×3 |
| Claude Code on the web | I21 | partly evidenced | Source | π Moderate · S1·R5·A1 ×3 |
| Claude Code, on your own machine | I21 | partly evidenced | Source | π Moderate · S1·R5·A1 ×3 |
| Claude Code, on your own machine | I22 | evidenced | Source | π Moderate · S1·R5·A1 ×3 |
| GetVocal AI agents | I21 | partly evidenced | "Each node defines what data the AI can access" Source | π Moderate · S1·R5·A1 ×3 |
| GitHub Copilot cloud agent | I21 | partly evidenced | "It can only push to a single branch" Source | π Moderate · S1·R5·A1 ×3 |
| GitHub Copilot cloud agent | I22 | evidenced | Source | π Moderate · S1·R5·A1 ×3 |
| Hackian | I22 | evidenced | "technologies and test types that Hackian does not currently handle on its own" Source | π Moderate · S1·R5·A1 ×3 |
| Junie | I21 | partly evidenced | "There are five types of actions that can be allowed with the" Source | π Moderate · S1·R5·A1 ×3 |
| Junie | I22 | evidenced | "Only run Junie non-interactively in projects you trust." Source | π Moderate · S1·R5·A1 ×3 |
| Manus | I21 | partly evidenced | "decides which local actions Manus runs on its own and which need approval" Source | π Moderate · S1·R5·A1 ×3 |
| Manus | I22 | evidenced | "Complex interactions, such as drag-and-drop and multi-step forms, may not work perfectly." Source | π Moderate · S1·R5·A1 ×3 |
| Noxus AI Workers | I21 | partly evidenced | "Select which actions the agent can call" Source | π Moderate · S1·R5·A1 ×3 |
| Noxus AI Workers | I22 | evidenced | "Web research is slower and less reliable than knowledge bases." Source | π Moderate · S1·R5·A1 ×3 |
| OutSystems Agent Workbench | I21 | partly evidenced | "You grant the agent the ability to use these actions to perform its tasks." Source | π Moderate · S1·R5·A1 ×3 |
| OutSystems Agent Workbench | I22 | evidenced | "Prompts or responses in other languages can bypass safety filters." Source | π Moderate · S1·R5·A1 ×3 |
| Replit Agent | I21 | partly evidenced | "Agent can read and write to them, but it doesn’t have access by default" Source | π Moderate · S1·R5·A1 ×3 |
| Replit Agent | I22 | evidenced | "Agent is looping or repeating the same steps" Source | π Moderate · S1·R5·A1 ×3 |
| Spinnable AI employees | I21 | partly evidenced | "A connection under Settings > Tools does not grant any worker access by itself." Source | π Moderate · S1·R5·A1 ×3 |
| Spinnable AI employees | I22 | evidenced | "Your worker initiates a task or multi-step workflow, but stops mid-way" Source | π Moderate · S1·R5·A1 ×3 |
| Visor.ai AI agents | I22 | evidenced | "The models generate each answer in the moment and can occasionally fail." Source | π Moderate · S1·R5·A1 ×3 |
Touched never means complies.
Answer it directly
A company can answer this line item itself, in its own file at /.well-known/fact-ray.json, with the node gamma-ray:provision/nist.MAP-2.2 (how). A company that already assesses itself against NIST AI RMF can reuse that work.