Self-Improving Business Skill#

Log business administration learnings, operational issues, and governance improvements to markdown files for continuous improvement. This skill focuses on business administration operations, cross-functional coordination, process governance, KPI tracking, execution risk, policy consistency, stakeholder communication, and operational decision quality.

CRITICAL: Safety Posture (Reminder-Only)#

This skill is documentation and reminder only.

It does not execute or authorize:

• spending approvals
• vendor commitments
• payroll actions
• legal actions
• procurement commitments
• policy sign-offs
• executive approvals

Use this skill to capture structured findings and recommendations. Human owners remain responsible for all high-impact business decisions.

Always require explicit human approval for high-impact decisions, including:

• budget reallocations
• vendor selection or termination
• policy exceptions
• operating model changes
• governance escalations
• staffing or compensation effects

First-Use Initialisation#

Before logging anything, ensure the .learnings/ directory and files exist in the project or workspace root. If any are missing, create them:

bash

mkdir -p .learnings
[ -f .learnings/LEARNINGS.md ] || printf "# Business Learnings\n\nBusiness administration learnings across process, governance, KPI tracking, and decision quality.\n\n**Categories**: process_breakdown | policy_gap | kpi_misalignment | decision_latency | handoff_failure | stakeholder_misalignment | documentation_drift | compliance_oversight\n**Areas**: planning | execution | governance | reporting | operations | budgeting | procurement | vendor_management | risk_management | communication\n\n---\n" > .learnings/LEARNINGS.md
[ -f .learnings/BUSINESS_ISSUES.md ] || printf "# Business Issues Log\n\nOperational execution issues, policy inconsistencies, KPI risks, and governance concerns.\n\n---\n" > .learnings/BUSINESS_ISSUES.md
[ -f .learnings/FEATURE_REQUESTS.md ] || printf "# Business Operations Feature Requests\n\nBusiness administration automation and process-improvement requests.\n\n---\n" > .learnings/FEATURE_REQUESTS.md

Never overwrite existing files. This is a no-op if .learnings/ is already initialised.

Do not treat this skill as an approval engine. It is a structured learning and reminder system only.

If you want automatic reminders, use the opt-in workflow in Hook Integration.

Quick Reference#

SituationActionMissed SLA in operational workflowLog to .learnings/BUSINESS_ISSUES.md with decision_latency or process_breakdownKPI value disputed due to conflicting definitionsLog to .learnings/LEARNINGS.md with kpi_misalignmentPolicy exists but team did not follow itLog to .learnings/BUSINESS_ISSUES.md with policy_gapHandoff failed between departmentsLog to .learnings/BUSINESS_ISSUES.md with handoff_failureStakeholder expectations diverge from execution planLog to .learnings/LEARNINGS.md with stakeholder_misalignmentDashboard differs from source-of-truth reportLog to .learnings/LEARNINGS.md with documentation_driftCompliance or audit requirement missed in workflowLog to .learnings/BUSINESS_ISSUES.md with compliance_oversightSimilar entry already existsAdd **See Also**, increase recurrence metadataReusable operating guidance emergesPromote to playbook/checklist/KPI registry/RACI/cadence artifact

OpenClaw Setup (Recommended)#

OpenClaw is the primary platform for this skill. It uses workspace-based prompt injection with automatic skill loading.

Installation#

Via ClawdHub (recommended):

bash

clawdhub install self-improving-business

Manual:

bash

git clone https://github.com/jose-compu/self-improving-business.git ~/.openclaw/skills/self-improving-business

Manual clone is optional. Check out a reviewed tag or commit before use. Do not install an unpinned default branch into a shared environment.

Create Learning Files#

bash

mkdir -p ~/.openclaw/workspace/.learnings

Then create the log files (or copy from assets/):

• LEARNINGS.md - reusable business learnings and governance improvements
• BUSINESS_ISSUES.md - execution problems, policy misses, and risk items
• FEATURE_REQUESTS.md - requested capabilities for administration workflows

Promotion Targets#

When business patterns are broadly applicable, promote them:

Learning TypePromote ToExampleRepeatable cross-team processProcess playbooks"Procurement intake triage in 2 business days"Governance controls and checksGovernance checklists"Quarterly policy attestation for budget owners"KPI definition clarityKPI definition registry"On-time delivery = committed date at order acceptance"Ownership ambiguity resolutionRACI matrix updates"Vendor SLA exceptions accountable by Operations Director"Execution rhythm improvementsOperating cadences"Weekly risk triage + monthly KPI review forum"Agent collaboration patternAGENTS.md"Route budgeting conflicts to governance triage agent"

Optional: Enable Hook#

Opt-in and workspace-local. Do not copy this hook to ~/.openclaw/hooks/ (that path is user-global). Skip it unless you want reminders in this workspace:

bash

mkdir -p .openclaw/hooks
cp -r hooks/openclaw .openclaw/hooks/self-improving-business

See references/openclaw-integration.md for complete details.

---

Generic Setup (Other Agents)#

For Claude Code, Codex, Copilot, or other agents, create .learnings/ in the project or workspace root:

bash

mkdir -p .learnings

Create the files inline using the headers shown above.

Agent file reminder#

Add a short note in AGENTS.md or CLAUDE.md to log LRN/BUS/FEAT entries and keep this workflow reminder-only with human approval for high-impact decisions.

Logging Format#

Learning Entry [LRN-YYYYMMDD-XXX]#

Append to .learnings/LEARNINGS.md:

markdown

## [LRN-YYYYMMDD-XXX] category

**Logged**: ISO-8601 timestamp
**Priority**: low | medium | high | critical
**Status**: pending
**Area**: planning | execution | governance | reporting | operations | budgeting | procurement | vendor_management | risk_management | communication

### Summary
One-line description of the business learning

### Details
Context of the finding: what happened, which teams were involved, where the process,
policy, KPI definition, or communication failed, and what improved outcome is expected.

### Recommended Action
Specific process correction, governance control, KPI definition update, or cadence change.
Reminder-only: this does not execute approvals.

### Metadata
- Source: meeting_review | retrospective | dashboard_audit | vendor_sync | finance_review | incident_postmortem
- Stakeholders: operations | finance | procurement | legal | hr | product | sales
- Related Files: path/to/process.md, path/to/report.csv
- Tags: tag1, tag2
- See Also: LRN-20260413-001
- Pattern-Key: ops.handoff_failure | kpi.definition_mismatch (optional)
- Recurrence-Count: 1 (optional)
- First-Seen: 2026-04-13 (optional)
- Last-Seen: 2026-04-13 (optional)

---

Categories for learnings:

CategoryUse Whenprocess_breakdownA repeatable business process fails due to missing or weak controlspolicy_gapExisting policy is absent, unclear, or not operationalizedkpi_misalignmentKPI meaning, source, or target differs across functionsdecision_latencyDecision gates are consistently delayed and impact executionhandoff_failureWork transfer between teams causes errors or delaysstakeholder_misalignmentStakeholders disagree on goals, scope, or prioritiesdocumentation_driftDocumentation diverges from actual process or system behaviorcompliance_oversightInternal or external compliance expectations are not met

Business Issue Entry [BUS-YYYYMMDD-XXX]#

Append to .learnings/BUSINESS_ISSUES.md:

markdown

## [BUS-YYYYMMDD-XXX] issue_type

**Logged**: ISO-8601 timestamp
**Priority**: critical | high | medium | low
**Status**: pending
**Area**: planning | execution | governance | reporting | operations | budgeting | procurement | vendor_management | risk_management | communication
**Severity**: critical | high | medium | low

### Summary
Brief description of the business issue

### Issue Details
What happened, why it matters, where dependencies broke, and what operational risk emerged.
Keep this as a documented finding and recommendation; this skill does not execute decisions.

### Impact Assessment
- Affected functions and owners
- KPI or SLA impact estimate
- Cost or budget impact range
- Governance or compliance implication

### Recommended Action
- Immediate containment steps
- Owner + escalation path
- Long-term process hardening
- Human approval required for any high-impact decisions

### Timeline
- **Identified**: ISO-8601
- **Escalated**: ISO-8601 (if applicable)
- **Resolved**: ISO-8601 (if applicable)

### Metadata
- Trigger: sla_miss | approval_delay | raci_conflict | kpi_variance | budget_alert | handoff_block | policy_missing | audit_finding | dependency_blocked
- Stakeholders: operations | finance | procurement | vendor_management | compliance | leadership
- Related Files: path/to/doc
- See Also: BUS-20260413-001

---

Feature Request Entry [FEAT-YYYYMMDD-XXX]#

Append to .learnings/FEATURE_REQUESTS.md:

markdown

## [FEAT-YYYYMMDD-XXX] capability_name

**Logged**: ISO-8601 timestamp
**Priority**: medium
**Status**: pending
**Area**: planning | execution | governance | reporting | operations | budgeting | procurement | vendor_management | risk_management | communication

### Requested Capability
What business administration capability is needed

### Business Justification
Why it improves speed, quality, policy consistency, or risk reduction

### Complexity Estimate
simple | medium | complex

### Suggested Implementation
How this could be built: workflow automation, dashboard guardrails, dependency tracker,
handoff checklist, policy linting, governance workflow reminders.

### Metadata
- Frequency: first_time | recurring
- Related Features: existing_tool_or_feature
- Benefit: cycle_time | decision_quality | risk_reduction | policy_consistency

---

ID Generation#

Format: TYPE-YYYYMMDD-XXX

• TYPE: LRN (learning), BUS (business issue), FEAT (feature request)
• YYYYMMDD: Current date
• XXX: Sequential number or random 3 chars (for example 001, A7B)

Examples: LRN-20260413-001, BUS-20260413-A3F, FEAT-20260413-002

Resolving Entries#

When an issue is resolved, update the entry:

1. Change **Status**: pending to **Status**: resolved
2. Add a resolution block after Metadata:

markdown

### Resolution
- **Resolved**: 2026-04-13T09:00:00Z
- **Action Taken**: Brief description of process or governance fix
- **Artifact Updated**: Playbook/checklist/KPI registry/RACI/cadence reference
- **Verified By**: operations review | finance review | governance committee | internal audit

Other status values:

• in_progress - actively being remediated
• wont_fix - risk accepted with documented rationale and human approver
• promoted - elevated to permanent business operating artifacts
• promoted_to_skill - extracted as reusable skill

Detection Triggers#

Automatically log when you encounter:

Execution Signals (learning or business issue):

• missed SLA on committed deliverable
• overdue approval at decision gate
• repeated dependency blocked status
• handoff without complete acceptance criteria

Governance Signals (learning or business issue):

• policy exists but team follows outdated practice
• control owner unknown or contested
• RACI accountability conflict between teams
• governance board receives contradictory status reports

KPI Signals (learning):

• KPI definition mismatch across dashboards
• variance unexplained beyond tolerance band
• lagging metrics reported as leading indicators

Budget and Procurement Signals (business issue):

• budget overrun warning without mitigation owner
• purchase request bypasses policy threshold
• vendor onboarding stalled at approval handoff

Compliance and Documentation Signals (learning):

• audit issue reveals missing process evidence
• runbook differs from actual operational behavior
• required approval record missing for material decision

Priority Guidelines#

PriorityWhen to UseBusiness ExamplescriticalMaterial operational disruption, uncontrolled budget risk, or compliance exposurePayroll-impacting process outage, major procurement exception without approval trailhighSLA misses, delayed governance decisions, or significant KPI varianceMissed month-end close dependency, recurring approval bottleneckmediumImportant process hardening or definition alignment workKPI glossary mismatch, outdated handoff checklistlowDocumentation cleanup and minor process optimizationNaming consistency updates, low-risk reporting cleanup

Area Tags#

Use these areas to classify learnings:

AreaScopeplanningplanning cycles, prioritization, resource alignmentexecutiontask flow, dependency management, delivery completiongovernanceapprovals, controls, operating model oversightreportingKPI reporting quality, cadence, dashboard trustoperationsday-to-day process reliability and service deliverybudgetingbudget planning, variance review, spend controlsprocurementpurchasing controls, approvals, sourcing processvendor_managementvendor onboarding, SLA adherence, performance reviewsrisk_managementrisk identification, escalation, mitigation trackingcommunicationcross-functional alignment, decision clarity, status clarity

Promoting to Permanent Business Standards#

When a learning is broadly applicable, promote it to business administration standards.

When to Promote#

• the same process breakdown recurs across teams
• policy gap repeatedly creates execution risk
• KPI definition issue causes repeated reporting conflicts
• decision latency appears in multiple planning or governance cycles
• handoff failures recur with shared root causes

Promotion Targets#

TargetWhat Belongs ThereProcess playbooksdocumented steps, guardrails, exception routingGovernance checklistsdecision controls, evidence requirements, escalation criteriaKPI definition registrymetric owners, formulas, data sources, review cadenceRACI matrix updatesrole clarity for accountable/responsible/consulted/informedOperating cadencesrecurring forums, review rituals, frequency and ownersCLAUDE.mdagent operating conventions for business contextAGENTS.mdcross-agent routing and workflow orchestration

How to Promote#

1. Distill the finding into concise and reusable operational guidance
2. Prepare a minimal patch for the correct target artifact
3. Show a reviewed diff and apply only after explicit user approval
4. Update original entry:
   • set **Status**: promoted
   • add **Promoted**: process_playbook (or other target)
5. Confirm human owner for adoption and enforcement

Promotion Examples#

Learning to governance checklist:

Overdue approvals repeatedly blocked procurement. Added checklist rule: "Any approval older than 48h triggers escalation to backup approver."

Learning to KPI registry:

Revenue-at-risk definition mismatched between sales and finance. Registered canonical formula and source system for both teams.

Recurring Pattern Detection#

If logging something similar to an existing entry:

1. Search first: rg "keyword" .learnings/
2. Link entries with **See Also**
3. Increment recurrence fields
4. Raise priority if impact is increasing
5. Trigger systemic fix proposal when recurrence exceeds threshold

Recurring business patterns usually indicate one or more root causes:

• missing process control
• weak policy communication
• unclear ownership model
• inconsistent KPI definition
• brittle cross-team handoff protocol
• insufficient governance cadence

Simplify/Harden Feed#

Keep incoming signals easy to process, and harden the learning feed against noise.

Simplify Rules#

• prefer one clear issue statement per entry
• include only evidence needed for a decision
• use standard categories and area tags
• avoid duplicating the same incident in multiple files
• route uncertain items to pending with explicit open questions

Harden Rules#

• require source attribution in metadata
• require owner and next action for high and critical
• require explicit approval note for wont_fix
• require recurrence tracking for repeated issues
• require status updates before review meetings

Periodic Review#

Review .learnings/ at operational checkpoints.

When to Review#

• before weekly operating review
• before monthly business review
• before quarterly planning cycle
• before budget review forums
• after major incident retrospectives
• before governance committee decisions

Quick Status Check#

bash

rg "Status\\*\\*: pending" .learnings/*.md | wc -l
rg "Priority\\*\\*: critical|Priority\\*\\*: high" .learnings/*.md
rg "Area\\*\\*: budgeting|Area\\*\\*: procurement" .learnings/*.md
rg "Category\\*\\*:|\\] process_breakdown|\\] decision_latency" .learnings/*.md

Review Actions#

• close resolved entries and capture verification evidence
• merge duplicates and strengthen canonical entries
• promote proven patterns to permanent artifacts
• add or adjust operating cadence for recurring issues

Hook Integration#

Enable automatic reminders through hooks. This is opt-in.

Hooks persist across sessions once installed. Keep them project-scoped. Do not install user-level or global hooks. Never use an empty matcher. PostToolUse inspects command output in-process; do not log raw output, secrets, or transcripts.

Quick Setup (Claude Code / Codex)#

Create .claude/settings.json:

json

{
  "hooks": {
    "UserPromptSubmit": [{
      "matcher": "kpi|revenue|margin|strategy|roadmap|pricing|forecast|RACI",
      "hooks": [{
        "type": "command",
        "command": "./skills/self-improving-business/scripts/activator.sh"
      }]
    }]
  }
}

This injects a business-administration reminder after matching prompts.

Advanced Setup (With Pattern Detection)#

json

{
  "hooks": {
    "UserPromptSubmit": [{
      "matcher": "kpi|revenue|margin|strategy|roadmap|pricing|forecast|RACI",
      "hooks": [{
        "type": "command",
        "command": "./skills/self-improving-business/scripts/activator.sh"
      }]
    }]
  }
}

Leave PostToolUse disabled. Do not attach error-detector.sh to every Bash call.

Available Hook Scripts#

ScriptHook TypePurposescripts/activator.shUserPromptSubmitreminds to evaluate business learningsscripts/error-detector.shPostToolUse (Bash)triggers on high-signal business patterns

See references/hooks-setup.md for details.

Automatic Skill Extraction#

Extracted skills are untrusted until a human reviews the generated SKILL.md. Do not keep or publish an extracted skill without explicit user approval.

When a learning becomes stable and reusable, extract it into a dedicated skill.

Skill Extraction Criteria#

CriterionDescriptionrecurringsame pattern appears in multiple teams or cyclesverifiedstatus is resolved with validated operational improvementnon-obviousrequired cross-functional reasoning or governance designbroadly applicableuseful outside a single incidentuser-flaggeduser requests conversion into a reusable skill

Extraction Workflow#

1. identify candidate entry

2. run helper: bash

   ./skills/self-improving-business/scripts/extract-skill.sh skill-name --dry-run
   ./skills/self-improving-business/scripts/extract-skill.sh skill-name
   
   

3. customize generated SKILL.md

4. update source entry status to promoted_to_skill

5. verify no transactional instructions were embedded

Multi-Agent Support#

AgentActivationDetectionClaude Codehooksautomatic via scriptsCodex CLIhooksautomatic via scriptsGitHub Copilotmanual guidancemanual reviewOpenClawworkspace injection + hookssession reminders

Regardless of agent, apply this workflow when you:

1. coordinate cross-functional delivery with dependencies
2. track KPI quality and reporting consistency
3. manage policy-governed approvals and exceptions
4. monitor budget, procurement, and vendor execution
5. document risk controls and escalation outcomes
6. resolve stakeholder alignment and communication gaps

Business Administration Best Practices#

1. keep ownership explicit and visible in each entry
2. separate signal from symptom using root-cause notes
3. normalize KPI definitions before target discussions
4. document decision deadlines and escalation rules
5. review RACI regularly for contested responsibilities
6. harden handoffs with acceptance criteria and evidence
7. tie governance controls to actual operating cadences
8. use consistent tags for reporting and searchability
9. require human approval for high-impact decisions
10. avoid turning reminders into hidden execution commands

Gitignore Options#

Keep learnings local when entries include sensitive business context:

gitignore

.learnings/

Hybrid option (track structure, ignore contents):

gitignore

.learnings/*.md
!.learnings/.gitkeep

Track learnings in-repo only when content is sanitized and approved for sharing.

Final Reminder#

This skill is a structured memory and improvement mechanism. It does not approve, execute, sign, spend, commit, or authorize business actions. Use it to improve clarity and decision quality, and require human approval for high-impact outcomes.

Stackability Contract (Standalone + Multi-Skill)#

This skill is standalone-compatible and stackable with other self-improving skills.

Namespaced Logging (recommended for 2+ skills)#

• Namespace for this skill: .learnings/business/
• Keep current standalone behavior if you prefer flat files.
• Optional shared index for all skills: .learnings/INDEX.md

Required Metadata#

Every new entry must include:

markdown

**Skill**: business

Hook Arbitration (when 2+ skills are enabled)#

• Use one dispatcher hook as the single entrypoint.
• Dispatcher responsibilities: route by matcher, dedupe repeated events, and rate-limit reminders.
• Suggested defaults: dedupe key = event + matcher + file + 5m_window; max 1 reminder per skill every 5 minutes.

Narrow Matcher Scope (business)#

Only trigger this skill automatically for business signals such as:

• kpi|revenue|margin|strategy|roadmap
• pricing|forecast|market shift|business case
• explicit business intent in user prompt

Cross-Skill Precedence#

When guidance conflicts, apply:

1. security
2. engineering
3. coding
4. ai
5. user-explicit domain skill
6. meta as tie-breaker

Ownership Rules#

• This skill writes only to .learnings/business/ in stackable mode.
• Do not read other skill folders, their SKILL.md files, or their log entries.
• Standalone mode writes to this project's .learnings/*.md log files only.
• Stackable mode writes only to the namespaced folder above and must not rewrite other skills' log entries.
• Promotion into AGENTS.md, SOUL.md, TOOLS.md, MEMORY.md, rules, hooks, or generated skills is not a logging write. Show a reviewed diff and apply only after explicit user approval.