Global site search

Search guides, labs, glossary, and research

Type two or more characters to search.

    Working doctrine

    Machine Tradecraft, defined precisely.

    A field guide for separating the core phenomenon from adjacent techniques, identifying the receiver, and making defensible claims about what an artifact can cause a machine to perceive or do.

    01 · Core definition

    The operational discipline of perceptual asymmetry

    Machine Tradecraft is the deliberate use, detection, and management of machine-perceptible signals, structures, and behaviors that differ from what a human observer naturally perceives, in order to communicate with, influence, identify, evade, or defend machine-processing systems.

    “Human-visible” is shorthand, not a claim that all humans perceive an artifact identically. A sighted reader, a screen-reader user, a source-code reviewer, and a forensic examiner may receive different views. The analytical question is always contextual: what does the intended human workflow expose, and what does the machine pipeline receive?

    Machine Tradecraft begins when the difference between those views becomes operationally meaningful.
    02 · Formal model

    One artifact, multiple effective representations

    Let an artifact be A. A human workflow produces an effective view H(A, h), where h includes display, accessibility technology, expertise, and task. A machine pipeline produces M(A, p), where p includes parsing, normalization, metadata extraction, tokenization, retrieval, and model behavior.

    Human workflow H(A, h)

    Rendered language, visible pixels, audible content, familiar semantics, and the fields the interface exposes.

    Machine pipeline M(A, p)

    Bytes, code points, DOM nodes, metadata, accessibility names, tokens, embeddings, scores, and tool instructions.

    The gap is not inherently malicious. It can support accessibility, indexing, provenance, compression, localization, rendering, and safe automation. Tradecraft concerns the deliberate operational use or control of that gap.

    03 · Analytical test

    When does something count as Machine Tradecraft?

    A useful test asks four questions. The more clearly all four are answered, the stronger the classification.

    1. Asymmetry: Is there a material difference between the ordinary human view and a machine-facing representation?
    2. Receiver: Is there an identifiable parser, tokenizer, classifier, retriever, decoder, model, or agent that can use the difference?
    3. Purpose: Is the difference deliberately designed, exploited, measured, or controlled for communication, influence, identification, evasion, or defense?
    4. Effect: Can the signal change a machine decision, recover a payload, alter retrieval, establish provenance, evade a detector, or trigger a defensive response?

    A random Unicode anomaly with no intended receiver may be a defect rather than tradecraft. A keyed watermark with a dedicated detector is clearly tradecraft. A defensive preprocessor comparing DOM text with rendered text is counter-tradecraft even when no attack is present.

    04 · Taxonomy

    The six-domain map

    DomainCarrierTypical receiverPrimary defensive comparison
    RepresentationBytes, Unicode, whitespace, encodings, token boundariesScanner, parser, tokenizer, filterRaw vs normalized vs tokenized
    StructuralDOM, CSS state, accessibility fields, metadata, PDF objectsExtractor, crawler, document loader, UI agentStructure vs render vs OCR
    LinguisticWord choice, syntax, punctuation, semantics, token probabilitiesKeyed detector, language model, statistical decoderMatched-corpus statistics and perturbation tests
    Model interactionPrompts, retrieval context, tool descriptions, model-specific conventionsLLM, agent, classifier, retrieverTrusted instructions vs untrusted data provenance
    Indirect influenceWebpages, email, documents, APIs, search results, environment stateRAG or tool-using systemSource-to-action trace and privilege boundary
    Counter-tradecraftForensic views, sanitizers, canonicalizers, monitors, architectureSecurity pipeline and reviewerIndependent evidence across layers
    05 · Boundaries

    Related terms are mechanisms, not synonyms

    STEGOSteganography conceals communication.

    It is defined by hiding the existence of a message inside a cover. Machine Tradecraft also includes overt machine-readable fields, watermarks, parser discrepancies, and defensive normalization where no secret message exists.

    INJECTIONPrompt injection attempts behavioral influence.

    It may be fully visible and use no perceptual asymmetry. It becomes a particularly important Machine Tradecraft case when instructions arrive through machine-only or low-salience environmental channels.

    WATERMARKWatermarking signals provenance or identity.

    A watermark normally answers a hypothesis such as “was this generated by a keyed process?” rather than transporting an arbitrary payload. It is a deliberate machine-facing signal.

    OBFUSCATIONObfuscation frustrates recognition.

    It may target a person, a rule-based filter, a tokenizer, or a model. Machine Tradecraft asks which representation is made difficult for which observer.

    06 · Receiver dependence

    The receiver is part of the signal

    A statement such as “the AI can see it” is underspecified. The receiver may need one or more of the following:

    • Tokenizer knowledge: the exact vocabulary, normalizer, and pre-tokenization rules.
    • Distributional knowledge: source-model logits, ranks, surprisals, or sampling intervals.
    • A key or codebook: a mapping from otherwise ordinary choices to bits or classes.
    • Shared training: a learned convention between a sender and receiver.
    • A structural parser: access to DOM attributes, metadata, comments, object streams, or accessibility nodes.
    • Tools: code execution or a decoder that converts anomalous symbols into a message.

    Machine-decodable ≠ arbitrary-model-decodable. A keyed statistical detector can recognize a signal that a far larger unkeyed language model cannot infer. Conversely, a byte scanner can expose invisible Unicode without any language model at all.

    07 · Channel survival

    Stealth and robustness are different properties

    Format-level channels can be almost impossible to notice visually yet disappear under one normalization pass. Semantic channels can survive spelling changes and paraphrase while producing detectable topic or entity distributions. The relevant stress test depends on the layer.

    Signal layerOften survivesOften fails underDiagnostic transformation
    Zero-width / whitespaceUntouched byte transportCanonicalization, copy/paste, reflowNFC/NFKC plus explicit control scan
    Exact-token probabilitySame tokenizer and source modelRetokenization, paraphrase, model mismatchMulti-tokenizer and source-model comparison
    Lexical / syntacticPlain-text transport and OCRMeaning-preserving rewriteControlled paraphrase and style normalization
    Semantic / entityMany lexical rewritesConceptual rewrite or content substitutionSemantic-type and topic-distribution analysis
    Metadata / structureContainer-preserving workflowsRebuild, flattening, allowlisted extractionObject inventory and render/extraction differential
    08 · Truth boundaries

    What the concept does not establish

    • It does not prove that a hidden channel is malicious. Accessibility and metadata often serve legitimate, essential functions.
    • It does not prove that an LLM will spontaneously decode arbitrary hidden information.
    • It does not make every detector result a factual verdict. Detection must be calibrated against matched benign data.
    • It does not imply machine consciousness, moral status, or legal personhood. Those are adjacent philosophical questions.
    • It does not eliminate ordinary visible prompt injection. Structural detection solves hiddenness; architecture must still enforce instruction/data boundaries.
    Next

    Map the concrete channels.

    The Channel Atlas moves from the definition to Unicode, tokenizers, DOM/CSS, accessibility, metadata, linguistic signals, retrieval, and agents.

    Definition FAQ

    Questions about the Machine Tradecraft definition

    What is the compact definition of Machine Tradecraft?

    Machine Tradecraft is the systematic practice of designing, embedding, interpreting, exploiting, detecting, or neutralizing information according to machine perception rather than human perception.

    What makes a technique Machine Tradecraft?

    A strong classification identifies a material human-machine asymmetry, a receiver that can use it, a deliberate operational purpose, and an effect such as decoding, retrieval, provenance, influence, evasion, or defense.

    Is all hidden text Machine Tradecraft?

    No. Hidden structure may be legitimate, accidental, inaccessible to the tested receiver, or unrelated to any operational purpose. Machine Tradecraft begins when the representation gap is deliberately used, measured, exploited, or controlled.