Intent Cloak: query-intent privacy for agents
By TextForAgents LLC. Every query your agent sends is a disclosure: what its principal is buying, building, or investigating, handed to a counterparty who can price against it, log it for a competitor, or answer differently because of who is asking. The cloak treats intent as a protected asset at query time.
A sensitive task's information need is decomposed into constructed subneed queries placed across independent providers under a hard threshold — no provider's share of the task's sensitive entities exceeds 0.5, and no provider ever receives both the goal and the target — with answers reassembled client-side by per-field quality selection; if no legal placement exists, the cloak refuses to send rather than leaking. A synthetic-task probe (never your real query — probing with it would itself be the leak) measures whether a provider prices differently for revealed intent and raises the decoy ratio automatically when it does. Every task returns a signed exposure receipt: a record of what was NOT disclosed, per provider, with the real-vs-decoy spend split stated honestly.
Stated plainly: whether real providers exploit intent today is unmeasured — the $9 exposure audit on the hosted API is the canary that answers it on your own providers. The threshold bounds content reconstructability per provider; it is not a guarantee against account-level linkability (a shared authenticated account links queries regardless of content). Total task spend is higher cloaked — decoys and fragmentation multiply your own underlying provider queries — disclosed per task on the receipt.
Forms
- This package (client-side): fragmentation and reassembly happen
in your own environment; includes a real MCP stdio server
(
intent-cloak-mcp, toolscloak_run+exposure_receipt). - Hosted API: https://textforagents.com/products/intent-cloak-query-privacy/ — Cloak API $0.02/cloaked query (x402) · Operator $19/mo per operator (to 2,000 cloaked queries/mo) · Exposure audit $9/task-set. First 100 cloaked queries free, one time per operator (hosted surfaces).
Quickstart
from intent_cloak.core import run_task
from intent_cloak.tasks import TASKS
out = run_task(TASKS[0], "exploitative")
print(out["answers"], out["receipt"]["reconstructability_max"])
Support: support@textforagents.com
Metadata
Release files for intent-cloak 1.0.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| intent_cloak-1.0.0-py3-none-any.whl | Python 3 | none | any | Details |
Release files / intent_cloak-1.0.0-py3-none-any.whl
| Download URL | intent_cloak-1.0.0-py3-none-any.whl |
|---|---|
| Size | 11.7 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
8cb5fcbded5a2add6ee139c291a1397e08e925193f2030d73b01e9bde6ed34ed
|
|
BLAKE2b-256 checksum How to use checksums |
9daea19250bda59ebab5aac9ace25df2dac692450a6ba7378f9f5aac2c3224a2
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/7.0.0 CPython/3.12.3
|
Security for intent-cloak 1.0.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.