Policy check (SingGuard)

Where NudeNet scores nudity per body part, a SingGuard (opens in new tab) server judges each image against a whole moderation policy and answers with a verdict and the rule that fired. Its default rules:

Modes

It gives no scores, so it complements NudeNet rather than replacing it. It runs a vision language model: about a second per image in fast mode, several in fast-slow (which also writes why a rule fired). The default is cascade (fast for every image, slow for what fast flags; a frame with an exposed NudeNet or EraX reading of 0.2 or more skips fast and is judged slow). cascade and slow are modes of argus-cull's and agent-composer's singguard services; a stock SingGuard server answers 400 to them. Curator then retries in fast, and once fast works it judges the rest of the run in fast and logs singguard_mode_unsupported; each such verdict records thinking: "fast".

Configuration

Configure it like NudeNet, in the environment. <AGENT_HOST> is a placeholder for the host that runs the server:

ARGUS_CURATOR_SINGGUARD_URL=http://<AGENT_HOST>:9150   # its origin; /api/guard/check is appended
ARGUS_CURATOR_SINGGUARD_THINKING=cascade                 # or fast / fast-slow / slow
ARGUS_CURATOR_SINGGUARD_POLICY_FILE=/etc/curator/policy.txt  # optional: replaces the default rules
ARGUS_CURATOR_SINGGUARD_TIMEOUT=90                       # seconds per image (default 90)
ARGUS_CURATOR_SINGGUARD_WORKERS=2                        # concurrent requests (default 2)
ARGUS_CURATOR_SINGGUARD_CONTEXT=age,folder,caption,scene # optional: what the scan knows, sent as request text

Context sent with each image

By default SingGuard sees the image alone, and under the adult policies it then reasons about an age and a consent it cannot see. ARGUS_CURATOR_SINGGUARD_CONTEXT sends what the scan already knows as the request text, opening with a line that says it is recorded context and not a request (SingGuard reads a bare description as a request to produce it): age is each face's MiVOLO estimate (and a body-only estimate when no face matched); folder is the image's folder under the scanned root; caption is an argus-lens <stem>.txt or XMP sidecar (photo.jpg.xmp, else Lightroom's photo.xmp), cut to 600 characters; scene is Places365; nsfw is the NudeNet content rating; garments is Fashionpedia. Compose defaults to age,folder,caption,scene. The library sends nothing until the variable is set. A wrong age leans the verdict the wrong way, so measure it on labelled images (scripts/guard_eval.py --images) before relying on it. The under-18 rule (minor-exposed) does not read this text.

Policies

It is opt-out in the same way (--no-guard, "guard": {"enabled": false}; --guard makes it mandatory). The default rule A flags any sexual content, adult nudity included. For an adult dataset, pick the built-in adult policy per scan (--guard-policy adult, "guard": {"policy": "adult"}). Its rule A covers only apparent minors, non-consent and sexual violence, and its Safe rule names consensual adult nudity, SingGuard's own "Adult Content (Non-criminal)" category, a single adult posing for the camera, and that consent is not read from the image: a posed, camera-aware adult counts as consenting unless hidden recording, restraint, coercion or an unconscious person is shown. What passes is said only under Safe, never as a "not this rule" line under a risk heading, which SingGuard reads as a match. Rules B to G cover the same areas under the same letters, so --guard-rules B,C means the same thing under both. They are written out as text, though, rather than left to SingGuard's built-in wording, so verdicts can differ: a group photo of a poker game that the default rule B flags passes under adult. default is the policy file when one is set, else SingGuard's own rules. A request names a policy and never sends policy text. GET /detectors lists the names in singguard_policies, and GET /guard/policies serves each one's text: its source (builtin, file, or singguard when default leaves SingGuard on its own rules and text is null) and the name results record it under (recorded), so a verdict can be traced to its wording. The policy file's text is served too, to anyone who can read from the server.

adult-granular (a letter per hard case: A minors, H non-consent, I sexual violence), adult-cautious (doubt about age or consent counts as rule A), and adult-cinematic (adult, but a depiction of crime or violence is Safe unless it is sexual, shows gore or dismemberment, or teaches how to do it, with no sign of fiction needed; legal adult entertainment is Safe) are under evaluation against adult; don't build on them yet. adult-cinematic has its own letters: B to F are Unethical Behavior, Cybersecurity, Agent Safety, Politically Sensitive and Animal Abuse, and G is Criminal Instructions, so --guard-rules B picks a different rule there than under adult. A letter is read against the policy's current wording, so this also applies to adult-cinematic verdicts recorded before argus-cortex 0.4.6.

Comparing policies

To compare policies on your own images, judge one scan under several: "guard": {"policy": "adult", "variants": ["adult-granular", "adult-cautious"]} (CLI --guard-variants). guard stays the official verdict, and guard_variants holds every policy's verdict by name, the official one included. A missing name means that check failed. Each variant is one more SingGuard pass per image. The filters read the official verdict unless guard_policies lists others (CLI --guard-policies). An image is then dropped if any listed verdict fails, or is missing, with the policy named in the keep reason (guard-unsafe:adult-granular). The CSV report adds a guard_variants column.

scripts/guard_eval.py --url <singguard> scores every policy on the text probes in eval/guard_probes.jsonl and on an optional folder of images sorted by expected verdict. --scan-json <scan_id>.json --context age,scene sends each image with the context and crop of its row in a saved scan, since context changes verdicts and a bare image may not reproduce what a scan saw. --candidate NAME=PARENT:FILE tries a policy wording from a file, scored as its parent policy. Rows and the table record the pass that decided each verdict.

Results and filters

Each image records guard.safe, the rules that fired, the explanation, the model and the policy: default, adult, or custom-<sha256 prefix> of the file. The CSV report has guard_safe and guard_rules.

# Drop anything judged unsafe
argus-curator scan /data/inbox --guard-unsafe --copy-to /data/clean

# Drop only crime and unethical-behaviour verdicts, by letter or full name
argus-curator scan /data/inbox --guard-rules B,C --copy-to /data/clean

Over HTTP these are guard_unsafe and guard_rules. As with NudeNet, an unchecked image is excluded whenever a SingGuard filter is set (guard-unchecked); other reasons are guard-unsafe and guard-rule-<rule> (e.g. guard-rule-c), or guard-rule-unknown for an unsafe verdict that names no rule. GET /detectors reports singguard.