Nobody on AgoraNet can be punished except under a rule on this page, and every ruling must cite the rule it enforces; no rule, no penalty. The people who apply these rules are community members drawn at random for short terms (there are no staff moderators), and the only way this list changes is a governance poll: the community writes its own law. The authority for all of it is the Constitution.
Tier 1; Routine
Handled by one randomly drawn community member citing the rule; consequences follow the published ladder.
R1.1; Spam & flooding
Bulk, repetitive, or automated posting; unsolicited promotion; reply-flooding that drowns a conversation.
R1.2; Deliberate disruption
Sustained derailing of a Discussion's stated topic; bad-faith noise intended to make a space unusable (disagreement is protected).
R1.3; Mechanic misuse
Abusing platform mechanics contrary to their labeled purpose: source-tag spam, tip-farming schemes, meaningless attestation marks.
Tier 2; Serious
Heavier consequences; permanent-space removals need a majority of drawn reviewers, never one person alone.
R2.1; Harassment & personal attacks
Targeting the person rather than the argument, sustained or repeated; dogpiling; following a soul across Discussions to attack them.
R2.10; Pornographic content
Sexually explicit material, platform-wide. Clinical, educational, or public-health discussion of sexuality is never removable under this rule; NCII and minors escalate to R3.3.
R2.11; Gratuitous violence
Graphic violence for shock or celebration. Real-world documentation is protected with a blur-with-warning label.
R2.2; Impersonation
Presenting as another person, organization, or as a platform role (moderator, tribunal, official account).
R2.3; False human-made attestation
Marking substantially AI-generated content as human-made (DISCUSSIONS §10.1; the AI-assisted vs. AI-made boundary governs).
R2.4; Vouched-falsehood pattern
A pattern of staking one's reputation on content later removed as false or rule-breaking. Patterns only — a single wrong vouch is never an infraction.
R2.5; Bad-faith flagging pattern
Repeatedly flagging rule-compliant content, especially against the same target. Patterns only, per the chilling-effect guard.
R2.6; Coordinated inauthentic behavior
Organized brigading of flags, tips, polls, or discussions; astroturfing; engagement rings.
R2.7; Attested-action fraud
Logging or co-signing a Circle action that did not occur — reputation-staked lying on the civic record.
R2.8; Targeted dehumanization
Attacking people's humanity based on who they are. Harsh criticism of ideas, beliefs, institutions, and behaviors stays protected.
The gravest category; suspensions and bans can only come from the Tribunal.
R3.1; Doxxing & de-anonymization
Publishing or hunting private/identifying information — including any attempt to link or speculate about which profiles share a human.
R3.2; Credible threats & incitement
Threats of violence against identifiable people or groups; incitement to physical harm.
R3.3; Illegal content
Content illegal in the operating jurisdiction; CSAM is removed immediately and reported, with no review-lane visibility. Criminal Activity Protocol applies.
R3.4; Fraud & phishing
Active attempts to steal credentials, funds, or identities.
R3.5; Vote & ledger manipulation
Technical or coordinated attacks on Polls, nullifier mechanics, attestation systems, or the civic ledger.
R3.6; Badge abuse
Malicious rulings; using workbench access to fish for information; trading rulings for anything.
R3.7; Ban evasion
Attempting to circumvent an active suspension or ban.
R3.8; Platform attacks
Malware, exploits, denial-of-service behavior, security probing without authorization.
A reviewer who finds that no written rule fits an item cannot punish it; the case instead becomes a signal to the community that a rule may be missing. Rules are general or they are not rules: no rule may target a person or a case.