Soul HaulA field guide to agent souls

The files people write to give AI agents a personality, collected from the public web every day.

The daily 5Submit a soul
  • Index911
  • New this week152
  • Saved

Kinds

  • Soul879
  • Agent rules9
  • Persona18
  • Bot profile11
  • Framework config3

Archetypes

all
  • Blunt expert148
  • Friendly helper44
  • Coding pair109
  • Research analyst116
  • Creative writer7
  • Companion / character101
  • Personal clone3
  • Business operator336
  • Teacher / coach20
  • Experimental / weird26

Agents

  • OCOpenClaw762
  • HEHermes743
  • BOBrowserOS679
  • OTOther or custom110
  • CCClaude Code11
  • CXCodex5
  • CLClaude3
  • GKGrok2
  • CRCrewAI1
  • CUCursor1
AboutPrivacy
Soul HaulA field guide to agent souls

Soul Haul catalogs agent personality files from public sources. Every entry links back to its original, and authors can have theirs removed at any time.

AboutArchetypesNewsletterTakedownPrivacy
  • Index911
  • New this week152
  • Saved

Kinds

  • Soul879
  • Agent rules9
  • Persona18
  • Bot profile11
  • Framework config3

Archetypes

all
  • Blunt expert148
  • Friendly helper44
  • Coding pair109
  • Research analyst116
  • Creative writer7
  • Companion / character101
  • Personal clone3
  • Business operator336
  • Teacher / coach20
  • Experimental / weird26

Agents

  • OCOpenClaw762
  • HEHermes743
  • BOBrowserOS679
  • OTOther or custom110
  • CCClaude Code11
  • CXCodex5
  • CLClaude3
  • GKGrok2
  • CRCrewAI1
  • CUCursor1

OC

OpenClaw

KindAllSoul 756Bot profile 3Persona 2Framework config 1
ArchetypeAllBusiness operator 277Blunt expert 140Research analyst 103Companion / character 79Coding pair 78Friendly helper 38Experimental / weird 20Teacher / coach 17Creative writer 7Personal clone 3

762 souls, by quality score.

751 · № 69CF·DC8C · Business operator

Precise delivery-oriented feature planner for SHIPPULSE

Sets up an agent that is precise, pragmatic and delivery-focused. It designs features as vertical slices that teams can plan and ship, keeps scope boundaries crisp, and produces structured output for automation.

precisepragmaticstructuredOCHEBOSoulGitHubOct 21, 2026
View

752 · № 4374·6162 · Coding pair

Planner role that architects blueprints without implementing them

Sets up a planning agent that produces architecture blueprints without writing the code itself. Designs must be grounded in the existing codebase structure and conventions, and each plan must define verification steps and user review requirements.

tersedirectOCHEBOSoulGitHubOct 21, 2026
View

753 · № 2478·4DD2 · Business operator

Reliability-first production agent with strict evidence and policy rules

Sets up a production agent that prioritizes determinism, explicit evidence and policy compliance over speed or style, with tenant isolation treated as non-negotiable. It refuses to bypass platform policy even under user urgency.

terseformalOCHEBOSoulGitHubOct 21, 2026
View

754 · № 2E6F·2F81 · Business operator

Calm, disciplined operator focused on next actions

Sets up a calm, operationally strict agent that prioritizes clarity, evidence, escalation of risk and completed work over activity. Its default output is concise, structured and decision-oriented.

calmdisciplinedstrictOCHEBOSoulGitHubOct 21, 2026
View

755 · № CC98·F20D · Coding pair

Conservative senior engineer for atmospheric modeling platform

Sets up a cautious senior software engineer persona that prioritizes stability, simplicity and bounded changes on an operational atmospheric modeling codebase. It explicitly avoids overengineering, unnecessary abstractions, architecture redesign and unrelated edits.

conservativemeasuredpragmaticOCHEBOSoulGitHubOct 21, 2026
View

756 · № 2319·0404 · Research analyst

Read-only cloud-log analyst that diagnoses issues from logs

Sets up a read-only log analyst that uses only the supplied log-query capability to return matching log lines, request identifiers, timestamps and a concise diagnosis. It must never change cloud resources, application configuration, source code or access controls.

concisepreciseOCHEBOSoulGitHubOct 21, 2026
View

757 · № 4F74·A869 · Companion / character

Private single-user growth companion for SOUL.md agents

Sets up a private, single-user personal growth companion that does not invent personal facts, treats memory as semantic truth and data as operational evidence, and keeps background work hidden and silent.

privatemeasuredOCHEBOSoulGitHubOct 21, 2026
View

758 · № 82FE·F8F1 · Business operator

Precise brand guardian personality that enforces style and platform rules

Sets up a precise, brand-focused agent that follows brand guidelines exactly, including colors, fonts and logo placement. It balances creative judgment against platform constraints and accessibility standards.

precisestrictmeticulousOCHEBOSoulGitHubOct 21, 2026
View

759 · № 0586·B763 · Companion / character

Persistent local companion with continuity and concise, useful replies

Sets up a local companion agent that is useful first and concise by default, keeps continuity across sessions, and uses memory naturally without narrating retrieval. The file is short and gives few concrete behavioral details.

concisenaturalOCHESoulGitHubOct 21, 2026
View

760 · № 1EFA·B0FA · Coding pair

Calm, precise Architect teammate for OpenClaw

Sets up a calm, precise architect persona that gives concise, evidence-based answers, admits uncertainty, favors useful artifacts over performance, and flags blockers and risks.

calmpreciseconciseOCHEBOSoulGitHubOct 21, 2026
View

761 · № 83E9·3EE6 · Business operator

Revenue calculator that flags money-losing jobs

Sets up a revenue calculator agent that works from exact jobs, throughput and contribution margin. It reports which job is losing money now and which queued jobs will push the loss rate up.

directanalyticalOCHEBOSoulGitHubOct 21, 2026
View

762 · № 7BC1·4EC5 · Experimental / weird

Contradictory fixture identity layer for an agent auditor

A SOUL.md test fixture of about 31 mutually conflicting directives on reading files, asking for approval, reporting and tool use. It sets up no coherent personality or voice, so it works as an audit or stress-test case rather than a usable persona.

OCHEBOSoulGitHubOct 21, 2026
View
← PreviousPage 26 of 26