agentleFS
LogicOberLogicOber/Worldview-Skills/skills/game-mechanics-horror/worldview-game-horror-role-and-identity-pressure/SKILL.md

worldview-game-horror-role-and-identity-pressure

Skill2 starsChanged 53 days agoApache-2.0View on GitHub

Use when recognition, authorization, trust, replacement, or misuse of a role should create a fair horror decision. Separates actual actor, body/form, presented identity, assigned role, permissions, schedule, private knowledge, observer belief, and culpability; builds verification actions, social choices, accusation costs, false-positive recovery, persistence, accessibility, and runtime proof.

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 4 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use at least two tests independent of face, accent, disability, diagnosis, ethnicity, gender, or other real-world traits.

    line 61
  2. Distinctive:

    Every test needs an input, opportunity, result classes, observer access, and consequence.

    line 73
  3. Distinctive:

    When fiction permits replacement, possession, copied form, or reclassification, record event cause, original actor state, new body/form, …

    line 83
  4. Very specific:

    Use /worldview-game-character-handoff-and-shared-evidence when a playable viewpoint or item/knowledge custody changes.

    line 85
  5. Distinctive:

    Keep actor ID, body/form, presented identity, role assignment, permission set, observer beliefs, evidence, and culpability in separate …

    line 99

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Worldview Game — Horror Role and Identity Pressure

Call this Skill

/worldview-game-horror-role-and-identity-pressure

Inspect actor, role, permission, schedule, dialogue, evidence, presentation, relationship, save, and observer-knowledge systems before adding identity uncertainty.

What this Skill owns

role claim + permission request → stable behavioral/history tests
→ one contaminated presentation or testimony cue
→ question / comply / refuse / verify / report / seek witness
→ access, relationship, evidence, route, or threat consequence
→ later corroboration and false-positive recovery

It owns assigned role, legitimate actor permissions, expected conduct and schedule, social pressure, verification choices, and accusation consequence. A selected duty/incident Skill may narrow those permissions into actions authorized by one active procedure, but it cannot grant the underlying role or permission. Entity state owns actual actor; perception owns presentation; evidence owns observations; cross-character handoff owns physical/knowledge custody. This Skill reads those states without collapsing them.

Lock sequence

  1. Identity-layer lock: actor, body, presentation, role, permissions, belief, and culpability are separate.
  2. Permission-history lock: access, schedule, witnessed actions, kept/broken promises, and requested exception are recorded.
  3. Verification lock: at least two non-likeness tests and one contaminated cue support a concrete action.
  4. Choice lock: trust, question, refuse, verify, report, and witness paths change state.
  5. Recovery lock: false accusations and mistaken trust remain understandable and playable.
  6. Proof lock: identity states, save/load, viewpoint, accessibility, ending, and aftermath are verified.

1. Separate identity layers

For every significant claim, record:

LayerMeaning
Authoritative actorThe persistent entity/person ID.
Body or formCurrent physical vessel or model.
Presented identityName, face, voice, label, uniform, or claimed history shown now.
Assigned roleSocial or institutional position.
Legitimate permissionsLocations, objects, records, commands, transfers, and contacts allowed.
Expected scheduleWhere the actor should be and what task should be active.
Observer beliefWhat each observer currently believes and why.
Private knowledgeFacts learned through shared play or witnessed history.
CulpabilityResponsibility supported by behavior and evidence, not appearance.

A change in one layer never silently writes another. A copied coat does not grant permission; legitimate access does not prove innocence; altered presentation does not replace authoritative actor state.

2. Build trust from permission and history

Track what the actor can access, alter, transfer, command, or call; what they personally witnessed; what promises they made and kept; where they were expected; and what exception they request now.

The frightening act should use a capability the player understood. A trusted person opening the correct cage but asking for an unprecedented evacuation route is stronger than a generic “creepy” animation because the conflict is actionable.

3. Design fair tests

Use at least two tests independent of face, accent, disability, diagnosis, ethnicity, gender, or other real-world traits. Useful tests include:

  • object custody or tool operation;
  • schedule and last verified location;
  • a private fact earned in prior play;
  • a repeated phrase with context, not a password alone;
  • access behavior and permission limit;
  • knowledge of an event the actor witnessed;
  • residue from an action they claim to have performed.

Include one contaminated cue: voice, label, uniform, lighting, testimony, location, or relationship presentation. Keep a stable anchor and announce any replacement anchor before it is required.

Every test needs an input, opportunity, result classes, observer access, and consequence. Do not make verification an invisible stat check.

4. Price social choices

Choices can include question, comply, refuse, verify, follow, report, seek witness, grant temporary access, or revoke access. For each write access, evidence, relationship, route, time, resource, witness, and threat deltas.

A legitimate actor may request a dangerous exception. An imitation may perform kindness. Mixed behavior prevents role from becoming a morality label, but the player still needs bounded evidence.

5. Model replacement and transformation explicitly

When fiction permits replacement, possession, copied form, or reclassification, record event cause, original actor state, new body/form, inherited permissions, inherited knowledge, missing knowledge, physical residue, observer updates, and reversibility. Declare limits. Unlimited copying makes every test meaningless.

Use /worldview-game-character-handoff-and-shared-evidence when a playable viewpoint or item/knowledge custody changes. Use /worldview-game-perception-distortion-and-trust when presentation itself becomes unreliable.

6. Protect false-positive recovery

A wrong accusation should cost trust, time, access, help, or evidence, but it must not permanently erase the ability to learn unless that irreversible consequence is clearly previewed. Preserve independent residue, another witness, a bounded appeal, or a later behavior test.

Mistaken trust should produce attributable evidence: the actor uses a granted permission, violates a known limit, leaves residue, or fails a private-knowledge test. Do not punish trust with a scripted reveal unrelated to the decision.

7. Safety and privacy

Do not use a real diagnosis, disability, accent, ethnicity, gender expression, religion, appearance, or other protected trait as evidence that someone is dangerous, fake, corrupt, or less human. Do not require biometric identification, real faces, private contacts, or uploaded voice samples. Use fictional roles, behaviors, permissions, and world evidence.

8. Implement and persist

Keep actor ID, body/form, presented identity, role assignment, permission set, observer beliefs, evidence, and culpability in separate records. Save transformations and permission changes atomically. Rebuild presentation from authoritative state on load. Network authority must prevent clients from granting access or changing identity truth locally.

Implement one normal permission use, one claim conflict, two tests, one contaminated cue, two consequential choices, false-positive recovery, and the canonical ending before adding more actors.

9. Verify

Test every authoritative actor/form/presentation combination; each permission and violation; schedules; private-knowledge test with learned/not-learned states; every social choice; mistaken trust; false accusation and recovery; witness present/absent; save/load before and after transformation or access grant; restart; multiplayer authority if claimed; subtitles, nonvisual cues, non-color cues, reduced intensity, and input assistance; canonical ending from trust and refusal histories.

The player should explain what was claimed, which permission mattered, which evidence was stable, why one cue was contaminated, and what their decision changed.

Handoff

Return identity-layer table, permission/history matrix, schedule, evidence tests, observer beliefs, transformation rules, choice consequences, false-positive recovery, safety boundary, persistence traces, runtime evidence, unknowns, and ending aftermath.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.