Skip to main content

agent-sanitizer (Python client)

A thin Python bridge to the agent-sanitizer Node.js CLI. The sanitization logic has a single source of truth — the JavaScript in src/ — so this package shells out to the CLI rather than re-implementing it, giving a Python pipeline byte-identical verdicts with no second implementation to keep in sync.

Requirements

  • Node.js (>= 22) on PATH. The sanitizer is JavaScript; something has to run it. There is deliberately no pure-Python fallback.

That's it — pip install agent-sanitizer and, with Node available, you're ready to go. The wheel ships a self-contained, single-file build of the CLI (the src/ logic and its npm dependencies bundled into one .mjs at release time), so there is no separate JavaScript checkout to clone and no environment variable to set. The bundle is a versioned build artifact from src/, not a hand-maintained port, so it can't drift from the JS.

Optional: point at your own JS checkout

Set AGENT_SANITIZER_CLI to a checkout's bin/sanitize-cli.mjs to override the bundled CLI (e.g. to run against unreleased src/ changes). When the module is imported directly from a repo checkout, the source CLI is found automatically.

Usage

from agent_sanitizer import sanitize

result = sanitize("untrusted text", html=True)
print(result.cleaned, result.found, result.warnings)

See the package docstring for the full set of entry points (sanitize_text, classify_prompt, scan_instruction_files, clean_file, and the long-lived Sanitizer worker).

Secret redaction ([secrets] extra)

The base install is dependency-free. The optional secrets extra adds a pure-Python secret-redaction engine under agent_sanitizer.secrets — detect-secrets plus custom detectors, benign-value skipping, cross-line reassembly, PEM collapse, and exact-match redaction of caller-supplied env-var values. Unlike the sanitizer above it needs no Node.js; its only dependency is detect-secrets, pulled in by the extra:

pip install "agent-sanitizer[secrets]"

Every detect-secrets import lives inside this subpackage, so a plain import agent_sanitizer never touches it. The engine shares the parent package's invisible-character SSOT (agent_sanitizer.invisible) rather than forking it — a fork would be a silent security regression.

In-process

from agent_sanitizer.secrets import RedactorConfig, redact, redact_map

redacted, found = redact("aws_key = AKIAIOSFODNN7EXAMPLE")
# -> ("aws_key = [REDACTED: AWS Access Key]", ["AWS Access Key"])

# Pass provider/host secret values in — config is supplied, never discovered.
cfg = RedactorConfig(host_cred_vars={"GH_TOKEN": "ghp_realtokenvalue123"})
redact("tok=ghp_realtokenvalue123", cfg)  # -> ("tok=[REDACTED: GH_TOKEN]", ["GH_TOKEN"])

The rehydration map contract

redact_map(text, config) returns a lossless, two-way view:

{"text": "<redacted>",
 "pairs": [{"placeholder": "[REDACTED: …]", "original": "<secret>", "start": <int>}],
 "found": ["<type>", ...]}

Substituting each pair["original"] at its start reconstructs the input byte-for-byte. start is a Unicode code-point offset into text (the field is start, not start_offset). U+E000 / U+E001 are reserved placeholder sentinels: input already containing either is refused with {"unmappable": "input contains reserved sentinel characters"} (fail closed) rather than producing an ambiguous map.

Daemon vs. one-shot

detect-secrets caches a process-global secret_type → plugin mapping, so a fresh one-shot call must re-register the plugin set every time — slow under load. Two console scripts (installed with the extra) cover both modes:

  • agent-secret-redactor — one-shot: reads text on stdin, writes the redaction result as JSON on stdout. Re-registers plugins per call.
  • agent-secret-redactor-daemon <socket-path> — long-lived Unix-socket server that configures the plugin set once at startup (a warm-up scan primes the cache before it binds, so a bound socket means ready). Wire protocol both directions: a 4-byte big-endian length prefix then that many bytes of UTF-8 JSON. Request {"text", "map", "web_ingress", "env_secrets"}; response is the one-shot result, JSON null when nothing is redacted, or {"error": …} on a scan failure.

Both console scripts import the engine on startup, so they fail loud if the [secrets] extra (hence detect-secrets) is not installed — fail closed, never a silent no-op.

Versioning

This package is versioned in lockstep with the npm agent-sanitizer: each release publishes both at the same version from the same commit, and the wheel bundles src/ at exactly that version. So pip install agent-sanitizer==X.Y.Z and npm i agent-sanitizer@X.Y.Z are the same underlying logic.

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

agent_sanitizer-2.2.0.tar.gz (295.5 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

agent_sanitizer-2.2.0-py3-none-any.whl (304.5 kB view details)

Uploaded Python 3

File details

Details for the file agent_sanitizer-2.2.0.tar.gz.

File metadata

  • Download URL: agent_sanitizer-2.2.0.tar.gz
  • Upload date:
  • Size: 295.5 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for agent_sanitizer-2.2.0.tar.gz
Algorithm Hash digest
SHA256 a2adb25295088c61c3a8a8ab2e7fa7d291bf7cb50556e71c5630056786257e0e
MD5 d189e4b5bc73af3bb2d973269a7b1e14
BLAKE2b-256 9efcc1e67baf4fcb0596bff9a4f894729baba420a3076b2395877fe94ff7b829

See more details on using hashes here.

Provenance

The following attestation bundles were made for agent_sanitizer-2.2.0.tar.gz:

Publisher: auto-version.yaml on AlexanderMattTurner/agent-sanitizer

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file agent_sanitizer-2.2.0-py3-none-any.whl.

File metadata

  • Download URL: agent_sanitizer-2.2.0-py3-none-any.whl
  • Upload date:
  • Size: 304.5 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for agent_sanitizer-2.2.0-py3-none-any.whl
Algorithm Hash digest
SHA256 eb56063547c2f1e0b92a4d9ed9ab918726d421ab7ecd8a7e15baa8e08175cbf4
MD5 2f6f35d0c96dac3c6bcb0511f8184bbf
BLAKE2b-256 10f186c60f2515f85285eb55b28842c3927aeda5cbdee3bab0d2085a7ddbd14c

See more details on using hashes here.

Provenance

The following attestation bundles were made for agent_sanitizer-2.2.0-py3-none-any.whl:

Publisher: auto-version.yaml on AlexanderMattTurner/agent-sanitizer

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Release history Release notifications | RSS feed

2.57.5

2 files

2.57.4

2 files

2.57.3

2 files

2.57.2

2 files

2.57.1

2 files

2.57.0

2 files

2.56.1

2 files

2.56.0

2 files

2.55.0

2 files

2.54.2

2 files

2.54.1

2 files

2.54.0

2 files

2.53.0

2 files

2.52.0

2 files

2.51.0

2 files

2.50.0

2 files

2.49.2

2 files

2.49.1

2 files

2.49.0

2 files

2.48.6

2 files

2.48.5

2 files

2.48.4

2 files

2.48.3

2 files

2.48.2

2 files

2.48.1

2 files

2.48.0

2 files

2.47.15

2 files

2.47.14

2 files

2.47.13

2 files

2.47.12

2 files

2.47.11

2 files

2.47.10

2 files

2.47.9

2 files

2.47.8

2 files

2.47.7

2 files

2.47.6

2 files

2.47.5

2 files

2.47.4

2 files

2.47.3

2 files

2.47.2

2 files

2.47.1

2 files

2.47.0

2 files

2.46.0

2 files

2.45.9

2 files

2.45.8

2 files

2.45.7

2 files

2.45.6

2 files

2.45.5

2 files

2.45.4

2 files

2.45.3

2 files

2.45.2

2 files

2.45.1

2 files

2.45.0

2 files

2.44.1

2 files

2.44.0

2 files

2.43.12

2 files

2.43.11

2 files

2.43.10

2 files

2.43.9

2 files

2.43.8

2 files

2.43.7

2 files

2.43.6

2 files

2.43.3

2 files

2.43.2

2 files

2.43.1

2 files

2.43.0

2 files

2.41.4

2 files

2.41.3

2 files

2.41.2

2 files

2.41.1

2 files

2.41.0

2 files

2.40.5

2 files

2.40.4

2 files

2.40.3

2 files

2.40.2

2 files

2.40.1

2 files

2.40.0

2 files

2.34.12

2 files

2.34.11

2 files

2.34.10

2 files

2.34.8

2 files

2.34.7

2 files

2.34.6

2 files

2.34.5

2 files

2.34.4

2 files

2.34.3

2 files

2.34.2

2 files

2.34.1

2 files

2.34.0

2 files

2.33.1

2 files

2.33.0

2 files

2.32.3

2 files

2.32.2

2 files

2.32.1

2 files

2.32.0

2 files

2.31.6

2 files

2.31.5

2 files

2.31.4

2 files

2.31.3

2 files

2.31.2

2 files

2.31.1

2 files

2.31.0

2 files

2.30.0

2 files

2.29.1

2 files

2.29.0

2 files

2.28.3

2 files

2.28.2

2 files

2.28.1

2 files

2.28.0

2 files

2.27.0

2 files

2.26.3

2 files

2.26.2

2 files

2.26.0

2 files

2.25.0

2 files

2.24.2

2 files

2.24.1

2 files

2.24.0

2 files

2.23.2

2 files

2.23.1

2 files

2.23.0

2 files

2.21.1

2 files

2.21.0

2 files

2.20.2

2 files

2.20.1

2 files

2.20.0

2 files

2.19.8

2 files

2.19.6

2 files

2.19.4

2 files

2.19.2

2 files

2.19.1

2 files

2.19.0

2 files

2.18.1

2 files

2.18.0

2 files

2.17.1

2 files

2.17.0

2 files

2.16.0

2 files

2.15.0

2 files

2.14.14

2 files

2.14.13

2 files

2.14.11

2 files

2.14.10

2 files

2.14.9

2 files

2.14.8

2 files

2.14.6

2 files

2.14.4

2 files

2.14.3

2 files

2.14.2

2 files

2.14.1

2 files

2.14.0

2 files

2.13.0

2 files

2.12.0

2 files

2.11.0

2 files

2.10.1

2 files

2.10.0

2 files

2.9.1

2 files

2.9.0

2 files

2.8.0

2 files

2.7.2

2 files

2.7.1

2 files

2.7.0

2 files

2.6.0

2 files

2.5.0

2 files

2.4.1

2 files

2.4.0

2 files

2.3.0

2 files

2.2.2

2 files

2.2.1

2 files

This release

2.2.0 This release

2 files

2.1.0

2 files

2.0.3

2 files

2.0.0

2 files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page