Sabotage Diagnostics
The discipline of distinguishing genuine dissent from strategic sabotage — five behavioral signatures read at the pattern scale, calibrated against the severe cost of false positives.
Full Practice · Bond · Cooperating Under Bad Faith
Mechanism
Genuine dissent and strategic sabotage can look indistinguishable from outside, and that is the problem the tool exists to address. Both oppose prevailing positions. Both name problems. Both make institutions uncomfortable. The difference is not in any single moment of conduct. It is in the pattern across many moments, and the pattern is legible only over time.
No single signature is conclusive. The pattern over time is the diagnostic.The tool reads behavior against five signatures.
Strongest vs. weakest version. Genuine dissent typically engages the strongest version of what it opposes. The dissenter wants to win the argument on its merits, which requires defeating the position at its strongest, not its weakest. Strategic sabotage typically engages the weakest version. The aim is not to win the argument but to disqualify the position by associating it with its worst defenders, its weakest formulations, or its most embarrassing examples. The signature is not the choice in any one engagement — sometimes the strongest version is genuinely unavailable, sometimes the weakest version is what is on the table. The signature is which version the actor reliably engages over many encounters.
Alternatives vs. fragmentation. Genuine dissent typically opposes one thing in service of something else: a better policy, a more accurate framing, a stronger institution, a more honest practice. There is something the dissenter is for, and the opposition is in service of building it. Strategic sabotage typically opposes without building. The aim is fragmentation rather than replacement. Each individual critique may be sharp, but across the arc no alternative ever coheres into a constructive proposal the actor would commit to.
Process acceptance vs. process attack. Genuine dissent typically accepts the process even when it produces unwanted outcomes. The dissenter believes the process is the right way to resolve disagreement and stays inside it. Strategic sabotage typically attacks the process when the process produces an unwanted outcome. Commitment to procedure holds for as long as procedure delivers the wanted result; when it does not, the procedure becomes the next target.
Consistency vs. audience-shifting. Genuine dissent typically holds the same substantive position in front of allies, opponents, and uncommitted audiences alike. The dissenter is making an argument they believe regardless of who is in the room. Strategic sabotage typically shifts with the audience. The substantive claim becomes whatever advances the strategic position in front of the current audience, even when the claims across audiences are mutually inconsistent.
Evidence vs. strategic advantage. Genuine dissent typically responds to evidence: when new information emerges, the position adjusts in proportion to what the information warrants. Strategic sabotage typically responds to strategic advantage: positions shift when shifting them improves the actor's position, even when no new evidence has emerged. The selection rule underneath the apparent reasoning is the diagnostic.
The signatures are not deterministic. Any one of them, in any one engagement, can be accounted for by good-faith explanations: the strongest version was unavailable, no alternative had yet been worked out, the process actually was flawed in this case, the audience required different emphasis, the strategic logic happened to coincide with the evidence. Each signature is ambiguous on its own. The pattern across many engagements, over months or years, is what carries information that no single signature carries.
The tool's deliberately high threshold is calibrated against a specific cost: the cost of false positives is severe. Labeling dissent as sabotage is among the oldest Control moves in institutional history. Every regime, every captured organization, every movement drifting toward closure has at some point used the language of "disruption" or "bad faith" or "not really one of us" to suppress critics who were operating in good faith. The tool's structural commitment is to err toward including dissent rather than excluding it, even at the cost of occasionally missing real sabotage early. The asymmetry of costs is what justifies the asymmetry of thresholds.
Practice
The diagnostic question is: "Is the pattern over time consistent with genuine disagreement, or with strategic disruption?"
The question is about pattern, not about person. It runs on observable behavior over a sufficiently long arc to surface what no single observation can show. Three practices operationalize it.
The Signature Audit. Run the five signatures against the actor's actual behavior across a window of multiple encounters. Record what the actor did — which version of opposing positions they engaged, whether they offered alternatives, what they did when losing inside the framework's process, whether their position shifted with audience, whether updates tracked evidence or strategic position. The audit is structured and external rather than impressionistic and internal. The point is to see what is actually there rather than what feels true, because the impressionistic reading is exactly where the tool can go wrong by reading discomfort with dissent as evidence of sabotage.
The Temporal Threshold. Refuse to act on findings from fewer than several independent observations over a window of several months. The exact threshold is judgment, not formula: enough observations to establish a pattern, enough time to rule out coincidence with a single difficult period in the actor's life. The discipline is to recognize that a single difficult engagement, a single sharp post, a single procedural objection produces noise rather than signal. The pattern requires accumulation, and the accumulation requires time the diagnostic is deliberately calibrated to require.
The Charitable-Interpretation Check. For each negative finding the audit produces, ask whether a good-faith interpretation accounts for the observation equally well or better. A reliably-weakest-version engagement might be a reliably-strategic actor, or it might be a critic who has not encountered the strongest versions because the institution surrounds them with weaker advocates. A reliably-procedure-attacking pattern might be strategic sabotage, or it might be a critic who has good reason to believe the procedure is itself the problem. The check is not a rationalization, it is a discipline: holding the negative finding against the strongest charitable reading and only counting the finding if the charitable reading does not account for it adequately. This is where most false positives are caught in advance.
The practices share the structure of slow judgment. None of them produces fast verdicts. The slowness is the discipline, because the cost of false positives is severe enough that fast verdicts are the failure mode the tool exists to prevent.
In the Wild
A political movement adopted reformist language in a national institution. Members attended every meeting and followed every procedure. When critics raised objections through the established processes, the movement used its majority to block them. When critics proposed procedural changes to restore balance, the movement invoked the sanctity of process. When the process produced outcomes the movement did not want, the movement attacked the process. Each move individually was defensible. The pattern across two years showed every signature: the engagement was reliably with the weakest version of opposing positions, no constructive alternative was ever committed to, the procedural commitment held only when procedure produced the wanted outcome, the substantive claims shifted with audience, and updates tracked the strategic position rather than the evidence. The diagnostic was clear at the pattern scale. It had been ambiguous at every single moment along the way.
A scholar engaged in years of public disputes inside their discipline. The disputes were sharp. Several colleagues at various points raised the possibility that the scholar was operating in bad faith. The Signature Audit produced a different reading. The scholar reliably engaged the strongest available formulations of opposing positions. The scholar offered concrete alternative accounts that they committed to defending. The scholar accepted disciplinary processes even when those processes produced rulings against their work. The scholar's position was consistent across audiences. Updates tracked evidence, including some embarrassing concessions when the evidence required them. The audit's finding was that the discomfort the scholar produced was the discomfort of genuine intellectual challenge, not sabotage. The tool's value in that case was protective: it prevented the framework from mistaking legitimate critic for adversary.
An organization developed a chronic pattern of treating any internal criticism as bad faith. Critics were screened out through informal channels well before any formal exclusion was contemplated. The organization characterized this practice as a defense against sabotage. The diagnostic ran the other direction. The organization was applying the language of bad-faith detection asymmetrically, without the temporal threshold, without the charitable-interpretation check, and without any audit of its own behavior against the same signatures. The tool's reading was that the organization had drifted into the Control failure mode the tool is calibrated to refuse. The framework's commitment to dissent was being eroded by exactly the rhetoric of bad-faith detection that this tool exists to discipline.
The tool's primary work is not detection. It is calibration. A framework that has no diagnostic for sabotage is naive against actors who arrive to disrupt it. A framework that has an overactive diagnostic for sabotage is the captured institution suppressing its own critics. The tool's structural commitment is to the calibration that does not become either failure mode.
What the framework asks of the practitioner using the tool is patience. Run the signatures. Wait for the pattern. Apply the charitable-interpretation check. Refuse to act on findings until the temporal threshold is cleared. The patience is what keeps the tool from becoming what it was built to refuse, which is the institutional habit of mistaking discomfort with dissent for evidence of bad faith.
Lineage
The institutional-theory literature on legitimate dissent provides the structural ground. Albert Hirschman's Exit, Voice, and Loyalty (1970) named voice as a constructive contribution to institutional health distinct from defection, and the body of work that grew from Hirschman tracks the conditions under which institutions either welcome voice as a corrective signal or suppress it as a threat. Sabotage Diagnostics inherits the distinction between voice (which the framework requires) and behaviors that look like voice but are functioning as exit-via-disruption.
The political-science literature on entryism documents the historical pattern from the other direction. Organized factions joining institutions in order to reshape them from inside has a long historical record, with characteristic behaviors that this tool's signatures abstract from many specific instances.
The historical record on dissent suppression — across regimes, organizations, and movements — provides the calibration that justifies the tool's high threshold. The rhetoric of bad-faith detection has been used to suppress legitimate critics so frequently and across so many contexts that any diagnostic for sabotage has to be built with explicit safeguards against becoming the next instance. The tool's structural commitment to charitable interpretation, slow judgment, and pattern-over-incident is a direct response to that record.
Game-theoretic work on disruption versus reform contributes the formal underpinning of the alternative-offering signature: actors with constructive goals typically commit to specific alternatives because doing so binds them to defendable positions, while actors with disruption goals typically avoid commitment because commitment forecloses strategic flexibility.
The Codex's own Loyal Opposition tool in the broader registry institutionalizes dissent as service rather than betrayal. Sabotage Diagnostics is the calibration that makes Loyal Opposition operable inside frameworks that also have to defend against actors who use the institution of dissent as their cover.
Cross-references
Within the category. Sabotage Diagnostics operates under the lens of Adversarial Dynamics, which carries the two-adversary framing and the broader diagnostic question. The tool relates to The Exclusion Problem as the upstream input to its downstream work: the Exclusion Problem's first condition for justified exclusion — a sustained pattern of bad faith rather than a single incident — is the kind of finding Sabotage Diagnostics produces. A framework cannot reach The Exclusion Problem responsibly without having run Sabotage Diagnostics with discipline first.
Across the Workshop. Receiving Disagreement Well is the Bond category Sabotage Diagnostics most directly serves and most relies on. Sabotage Diagnostics' false-positive defense depends on the framework being competent at the practice of receiving genuine disagreement; if the framework cannot receive sharp critics well even when they are operating in good faith, the diagnostic will read every sharp critic as a sabotage signal because the framework has no functional alternative for handling them. The Range geometry runs both ways: the practice of receiving disagreement well makes the sabotage diagnostic credible, and the sabotage diagnostic makes the practice of receiving disagreement well sustainable against actors who would otherwise exploit it.
Limitations. Three worth naming. First, the diagnostic is fundamentally probabilistic. Even a well-applied audit can be wrong about a specific actor, and the tool's structural commitment to err on the side of inclusion means some sabotage will slip past the threshold. The tool cannot eliminate uncertainty; it can only improve the calibration. Second, the tool can be wielded by parties who themselves operate in bad faith — an asymmetric actor running the audit against legitimate critics is the chronic risk the tool's own safeguards (charitable-interpretation check, temporal threshold, audit of the auditor's own behavior) are calibrated against, but no procedure eliminates the risk entirely. Third, in institutional contexts where the tool is in active use, the tool itself becomes part of the field actors are operating in. A strategic actor who reads the tool will adjust their behavior to pass the audit. The tool's defense against this is the pattern-and-time scale at which the signatures operate, which cannot be fully gamed without dropping the strategic logic the tool exists to detect.