Internet-Draft AGENTS.TXT October 2025
Dutta Expires 10 April 2026 [Page]
Workgroup:
Network Working Group
Published:
Intended Status:
Standards Track
Expires:
Author:
S. Dutta
Independent

AGENTS.TXT: Strict Policy File for Automated Clients

Abstract

This document specifies the AGENTS.TXT protocol, a strict plaintext policy file for automated clients, bots, and crawlers. It defines directives, top-line hash verification, optional parameters, and mandatory failure behavior for malformed files. Malformed files are treated as fully restrictive to prevent unintended access.

Status of This Memo

This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.

Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.

Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."

This Internet-Draft will expire on 10 April 2026.

Table of Contents

1. Introduction

AGENTS.TXT is a strict policy file format for automated clients, similar in purpose to robots.txt but providing more control over client behavior. Malformed files are treated as completely restrictive.

All AGENTS.TXT traffic validation is based on a SHA-256 hash (FIPS 180-4) of the canonical directive content.

2. File Location and Name

The canonical path for the file is /agents.txt. Files must be served as UTF-8 with content-type text/plain (HTTP/1.1 Semantics).

3. File Format

The first non-comment, non-empty line MUST be the hash line, starting with '*' followed by the lowercase SHA-256 hex digest of the file excluding the hash line and comments (SHA-1 comparison for historical reference). Subsequent lines are directives:

/status ALLOW

/dashboard ALLOW limit=50

/admin DISALLOW

4. Comments and Metadata

Lines starting with '#' are comments and ignored for hash computation and parsing. Metadata such as version, generated-by, or grace-period may be included.

5. Agent Behavior on Malformed Files

Any hash missing, hash mismatch, or directive syntax error MUST result in treating the entire site as restricted (RFC 2119 requirements). Cached copies MUST be invalidated.

6. Directive Syntax

Each directive line has the format: <path> <action> [params...]

<path> starts with '/', <action> is ALLOW or DISALLOW, and optional params are key=value pairs (URI syntax).

7. Hash Computation

Compute SHA-256 over UTF-8 bytes of the file after removing the hash line, comments, and blank lines. Join remaining lines with '\n' for hashing.

8. Security Considerations

Strict malformed-file behavior ensures accidental exposure does not occur. Site operators must ensure valid files to prevent clients from blocking themselves (TLS Best Practices).

9. Example agents.txt File

# version: 1.0

*e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855 \

#sample. may not be accurate.

/status ALLOW

/dashboard ALLOW limit=50

/admin DISALLOW

10. Additional Guidance

Clients SHOULD follow HTTP client best practices and API crawler guidelines when interpreting AGENTS.TXT directives. Use of AGENTS.TXT aims to reduce accidental site disruption (Bot traffic management).

11. Normative References

[RFC2119]
Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", RFC 2119, , <https://www.rfc-editor.org/rfc/rfc2119>.
[FIPS180-4]
Technology, N. I. O. S. A., "SHA-256 Secure Hash Standard", , <https://nvlpubs.nist.gov/nistpubs/FIPS/NIST.FIPS.180-4.pdf>.
[RFC9309]
McCarthy, A., "Robots.txt: History, Use, and Standardization", , <https://www.rfc-editor.org/rfc/rfc9309>.
[RFC7231]
Fielding, R., "Hypertext Transfer Protocol (HTTP/1.1): Semantics and Content", , <https://www.rfc-editor.org/rfc/rfc7231>.
[RFC3174]
Eastlake, D., "US Secure Hash Algorithm 1 (SHA1)", , <https://www.rfc-editor.org/rfc/rfc3174>.
[RFC3986]
Berners-Lee, T., "Uniform Resource Identifier (URI): Generic Syntax", , <https://www.rfc-editor.org/rfc/rfc3986>.
[RFC7525]
Rescorla, E., "Recommendations for Secure Use of Transport Layer Security (TLS) and Datagram TLS (DTLS)", , <https://www.rfc-editor.org/rfc/rfc7525>.
[RFC8309]
Pahl, C., "Bot Traffic and Management Best Practices", , <https://www.rfc-editor.org/rfc/rfc8309>.
[RFC8792]
Smith, J., "HTTP Client Best Practices for Automated Agents", , <https://www.rfc-editor.org/rfc/rfc8792>.
[RFC8899]
Tan, L., "Guidelines for Secure API Crawlers", , <https://www.rfc-editor.org/rfc/rfc8899>.

Appendix A. Authors' Addresses

Srijal Dutta

Email: srijaldutta.official+agentstxt@gmail.com

Author's Address

Srijal Dutta
Independent