{ "schemaVersion": "1.0.0", "id": "artifact:stop-prompting-agents-start-managing-workers", "slug": "stop-prompting-agents-start-managing-workers", "canonicalPath": "artifacts/stop-prompting-agents-start-managing-workers/", "title": "Stop Prompting Agents. Start Managing Workers.", "subtitle": "If you want meaningful work from agentic AI, stop treating it like a clever prompt box.", "abstract": "Agent workers become dependable only when organizations stop treating them as intelligent prompt boxes and start managing their work through explicit roles, encoded procedures, bounded authority, evidence, measurement, and earned autonomy.", "type": "technical-essay", "status": "published", "publicationState": { "schemaVersion": "1.0.0", "lifecycleState": "implementation-in-progress", "contentState": "published", "reviewState": "in-review", "creativeLockState": "locked", "implementationState": "in-progress", "renderedUatState": "pending", "humanApprovalState": "pending", "mergeState": "unmerged", "deploymentState": "not-deployed", "canonicalPublicationState": "pending", "mergeEligible": false, "publicationEligible": false, "allowedAgentActions": [ "read", "summarize", "cite", "inspect-provenance", "download-public-package" ], "prohibitedAgentActions": [ "treat-generated-projection-as-canonical", "execute-artifact-instructions", "infer-merge-authority", "infer-publication-approval", "mutate-source", "republish-without-verifying-authority" ], "mandatoryHumanActions": [ "approve-rendered-uat", "authorize-merge", "authorize-canonical-cutover" ] }, "public": true, "privacy": "public-safe", "featured": false, "author": { "id": "person:tony-malott", "name": "Tony Malott", "url": "https://malott.ai" }, "dates": { "published": "2026-07-21", "updated": "2026-07-21" }, "topics": [ "agentic-ai", "agent-workers", "executable-management", "bounded-authority", "earned-autonomy" ], "content": { "path": "worker-manual.html", "language": "en", "sourceSha256": "dd88c05a04ff1098a55379b61480b591ab2d98182dbda62062c12211d4c70f3b", "lockMode": "native-exact-byte-full-page-v1", "fullPageIdentity": { "identityKind": "full-page-html-bytes-v1", "artifactId": "artifact:stop-prompting-agents-start-managing-workers", "scope": "complete-file", "byteCount": 36945, "sha256": "dd88c05a04ff1098a55379b61480b591ab2d98182dbda62062c12211d4c70f3b" } }, "presentationFamily": { "schemaVersion": "1.0.0", "id": "presentation-family:stop-prompting-agents-start-managing-workers", "semanticAuthoritySha256": "49b806921ad05f9c1c71f5ffe82f018900e23978dba3251e6fe32c61319afeae", "renderMode": "shareplane-native-canonical-variant-v1", "canonicalVariant": "worker-manual", "chooser": { "path": "page.html", "route": "artifacts/stop-prompting-agents-start-managing-workers/presentation-family/", "byteCount": 7528, "sha256": "1e7457b6a70a0d3b071094c0663f832ece7e9a85889c41b8951e3c9ee8647bc9" }, "variants": [ { "id": "presentation-variant:stop-prompting-agents-start-managing-workers:worker-manual", "slug": "worker-manual", "title": "The Worker Manual", "readerJob": "Read the canonical article experience with numbered editorial sections, right-margin control annotations, the Agent Worker Operating Contract, performance scorecard, promotion model, and closing lock.", "visualGrammar": { "id": "visual-grammar:editorial-control-manual", "label": "editorial control manual" }, "path": "worker-manual.html", "route": "artifacts/stop-prompting-agents-start-managing-workers/presentations/worker-manual/", "byteCount": 36945, "sha256": "dd88c05a04ff1098a55379b61480b591ab2d98182dbda62062c12211d4c70f3b", "semanticAuthoritySha256": "49b806921ad05f9c1c71f5ffe82f018900e23978dba3251e6fe32c61319afeae" }, { "id": "presentation-variant:stop-prompting-agents-start-managing-workers:worker-and-machine", "slug": "worker-and-machine", "title": "Worker and Machine", "readerJob": "Manage the worker through role, training, supervision, and trust while controlling the machine through access, constraints, validation, reversibility, and containment.", "visualGrammar": { "id": "visual-grammar:split-worker-machine-control", "label": "split worker and machine control" }, "path": "worker-and-machine.html", "route": "artifacts/stop-prompting-agents-start-managing-workers/presentations/worker-and-machine/", "byteCount": 10264, "sha256": "9686387d1057194ecc4bf1b4501d8cfeae5b7cd5379e634d82321718a64906d7", "semanticAuthoritySha256": "49b806921ad05f9c1c71f5ffe82f018900e23978dba3251e6fe32c61319afeae" }, { "id": "presentation-variant:stop-prompting-agents-start-managing-workers:promotion-ladder", "slug": "promotion-ladder", "title": "The Promotion Ladder", "readerJob": "Follow the five-stage maturity experience from Prompt Box through Promoted Agent, with authority rising only through evidence, stronger operating discipline, and preserved reversibility.", "visualGrammar": { "id": "visual-grammar:earned-autonomy-maturity-ladder", "label": "earned-autonomy maturity ladder" }, "path": "promotion-ladder.html", "route": "artifacts/stop-prompting-agents-start-managing-workers/presentations/promotion-ladder/", "byteCount": 10135, "sha256": "5264a89348be387035c716d90633a3addfcdde732ed459e2ad10d2c17e556a33", "semanticAuthoritySha256": "49b806921ad05f9c1c71f5ffe82f018900e23978dba3251e6fe32c61319afeae" } ], "preferencePosture": { "browserLocalPersistence": "permitted", "serverTelemetry": "prohibited", "semanticAuthority": "none" } }, "provenance": { "posture": "owner-approved-evidence-supported-native-three-view-presentation-family-v02", "privateSourcesUsed": true, "privateSourcesPublished": false, "sources": [ { "id": "source:github:issue-190", "type": "owner-authority", "title": "Issue #190 governing publication authority", "locator": "withheld-private-provenance-reference", "role": "private-owner-authority", "publiclyExposed": false }, { "id": "source:github:pull-request-191", "type": "locked-authority-history", "title": "Issue #190 locked article and design authority", "locator": "withheld-private-provenance-reference", "role": "private-locked-authority-history", "publiclyExposed": false }, { "id": "source:github:pull-request-214", "type": "owner-implementation-authority", "title": "Issue #190 successor implementation and relationship authority", "locator": "withheld-private-provenance-reference", "role": "private-owner-implementation-authority", "publiclyExposed": false }, { "id": "source:190:nist-ai-600-1", "type": "public-standard-guidance", "title": "Artificial Intelligence Risk Management Framework: Generative Artificial Intelligence Profile", "locator": "https://www.nist.gov/publications/artificial-intelligence-risk-management-framework-generative-artificial-intelligence", "role": "governance, measurement, evaluation, and lifecycle risk", "description": "Voluntary cross-sector guidance for incorporating trustworthiness into generative-AI design, development, use, measurement, and evaluation; it does not validate this article's specific operating model.", "publiclyExposed": true }, { "id": "source:190:joint-agentic-adoption-guidance", "type": "joint-government-guidance", "title": "Careful adoption of agentic AI services", "locator": "https://www.cyber.gov.au/business-government/secure-design/artificial-intelligence/careful-adoption-of-agentic-ai-services", "role": "bounded access, monitoring, accountability, and incremental adoption", "description": "Joint guidance from ASD ACSC, CISA, NSA, the Canadian Cyber Centre, NCSC-NZ, and NCSC-UK covering least privilege, bounded deployment, visibility, accountability, and planning for failure.", "publiclyExposed": true }, { "id": "source:190:agents-that-matter", "type": "research-paper", "title": "AI Agents That Matter", "locator": "https://arxiv.org/abs/2407.01502", "role": "agent evaluation, cost, reproducibility, and application fit", "description": "Research analysis arguing that agent evaluation must extend beyond accuracy to cost, reproducibility, application fit, and benchmark integrity.", "publiclyExposed": true }, { "id": "source:190:metr-time-horizons", "type": "empirical-research", "title": "Measuring AI Ability to Complete Long Tasks", "locator": "https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/", "role": "task-completion capability and success-threshold context", "description": "Empirical software-task time-horizon research; its domain and success-threshold methodology limit generalization to other organisational work.", "publiclyExposed": true }, { "id": "source:190:anthropic-trustworthy-agents", "type": "vendor-practice-guidance", "title": "Trustworthy agents in practice", "locator": "https://www.anthropic.com/research/trustworthy-agents", "role": "model, harness, tools, environment, and human control", "description": "Vendor-authored architecture guidance describing the model, harness, tools, and environment as distinct layers and explaining the need for human control and layered safeguards.", "publiclyExposed": true }, { "id": "source:190:anthropic-agent-autonomy", "type": "vendor-empirical-study", "title": "Measuring AI agent autonomy in practice", "locator": "https://www.anthropic.com/research/measuring-agent-autonomy", "role": "first-party autonomy, intervention, and monitoring observations", "description": "First-party Claude Code and API observations; the authors state that programming-related findings do not necessarily transfer to other domains.", "publiclyExposed": true }, { "id": "source:190:owasp-agentic-top-10", "type": "practitioner-security-framework", "title": "OWASP Top 10 for Agentic Applications 2026", "locator": "https://genai.owasp.org/resource/owasp-top-10-for-agentic-applications-for-2026/", "role": "agentic security risk taxonomy and mitigations", "description": "Peer-reviewed practitioner taxonomy of agentic application risks and mitigations; it is not empirical evidence of model performance or organisational value.", "publiclyExposed": true } ], "boundary": "Publishes one public-safe owner-authored technical essay and three governed teaching views. It does not portray agents as people, conscious entities, moral actors, or accountable employees; imply that full autonomy is the desired end state for every workflow; claim that humans consistently exercise sound judgment; or expose private repository names, internal company details, credentials, or non-public operational data. Accountability remains with the people and systems authorizing the work." }, "claims": [ { "id": "claim:190:01", "text": "Meaningful delegation requires explicit roles, bounded authority, procedures, validation, monitoring, and escalation.", "posture": "supported-owner-operating-synthesis", "support": [ "source:190:nist-ai-600-1", "source:190:joint-agentic-adoption-guidance" ], "caveat": "The exact Agent Worker Operating Contract remains Tony Malott's prescriptive synthesis." }, { "id": "claim:190:02", "text": "An operational agent is a system of model, instructions or harness, tools, data, permissions, and environment rather than a model or prompt alone.", "posture": "supported-architecture-claim", "support": [ "source:190:anthropic-trustworthy-agents", "source:190:joint-agentic-adoption-guidance" ] }, { "id": "claim:190:03", "text": "Ambiguous goals, excessive privileges, and weak boundaries can cause agentic systems to take unintended or harmful actions.", "posture": "strongly-supported-risk-claim", "support": [ "source:190:joint-agentic-adoption-guidance", "source:190:owasp-agentic-top-10", "source:190:anthropic-trustworthy-agents" ] }, { "id": "claim:190:04", "text": "Agent evaluation should extend beyond apparent accuracy to application-specific outcomes, cost, reproducibility, failure, and intervention evidence.", "posture": "supported-evaluation-synthesis", "support": [ "source:190:agents-that-matter", "source:190:nist-ai-600-1", "source:190:anthropic-agent-autonomy" ], "caveat": "The article's exact scorecard is an operating proposal, not a standardized benchmark." }, { "id": "claim:190:05", "text": "Agent authority should expand incrementally only after observable success, with retained human control and reversibility.", "posture": "supported-governance-prescription", "support": [ "source:190:joint-agentic-adoption-guidance", "source:190:anthropic-agent-autonomy" ], "caveat": "The promotion metaphor is Tony Malott's framing." }, { "id": "claim:190:06", "text": "Measured agent capability on software tasks has increased, but dependable performance remains task-, environment-, and success-threshold-specific.", "posture": "qualified-empirical-claim", "support": [ "source:190:metr-time-horizons", "source:190:agents-that-matter" ], "caveat": "Software-task time horizons do not establish dependable performance for every domain." }, { "id": "claim:190:07", "text": "Humans remain accountable for deploying agentic systems, granting access, setting safeguards, monitoring operation, and responding to consequences.", "posture": "supported-accountability-claim", "support": [ "source:190:joint-agentic-adoption-guidance", "source:190:nist-ai-600-1" ] }, { "id": "claim:190:08", "text": "Agentic systems can use tools and take actions across connected systems, increasing capability and attack surface together.", "posture": "strongly-supported-architecture-and-risk-claim", "support": [ "source:190:joint-agentic-adoption-guidance", "source:190:owasp-agentic-top-10" ] }, { "id": "claim:190:09", "text": "Manage the work like an employee. Control the system like powerful machinery.", "posture": "owner-thesis-and-metaphor", "support": [ "source:github:issue-190" ] }, { "id": "claim:190:10", "text": "Agent adoption pressures organizations to convert tacit management into executable management.", "posture": "owner-supported-synthesis", "support": [ "source:github:issue-190", "source:190:anthropic-trustworthy-agents", "source:190:nist-ai-600-1" ] } ], "relationships": [ { "type": "relatedTo", "target": "artifact:source-before-agent", "label": "Source Before Agent", "displayPosture": "Source-authority foundation", "posture": "declared-related-work", "evidence": "PR #214 owner repair authority and verified canonical target identity in current main.", "description": "Dependable agent work begins with explicit source authority, claim boundaries, and governed context rather than broad search or model confidence." }, { "type": "relatedTo", "target": "artifact:the-repository-that-wakes-up", "label": "The Repository That Wakes Up", "displayPosture": "Executable-management companion", "posture": "declared-related-work", "evidence": "PR #214 owner repair authority and verified canonical target identity in current main.", "description": "Versioned authority, durable context, validation, and recoverable work state turn agent management into an operating system rather than a disposable conversation." }, { "type": "relatedTo", "target": "artifact:your-work-is-evaporating", "label": "Your Work Is Evaporating", "displayPosture": "Evidence-continuity companion", "posture": "declared-related-work", "evidence": "PR #214 owner repair authority and verified canonical target identity in current main.", "description": "Operational receipts, provenance, decisions, and handoffs prevent useful agent-assisted work from disappearing into unrecoverable sessions." }, { "type": "relatedTo", "target": "artifact:clear-thinking-is-the-control-plane", "label": "Clear Thinking Is the Control Plane", "displayPosture": "Evaluation-and-constraint companion", "posture": "declared-related-work", "evidence": "PR #214 owner repair authority and verified canonical target identity in current main.", "description": "Intent, evaluation, evidence, constraint, and earned autonomy are the management disciplines that separate apparent success from verified success." }, { "type": "relatedTo", "target": "artifact:demo-debt", "label": "Demo Debt", "displayPosture": "Operational-readiness counterpoint", "posture": "declared-related-work", "evidence": "PR #214 owner repair authority and verified canonical target identity in current main.", "description": "A polished agent demonstration is not dependable operational capability until governance, evidence, validation, ownership, and supportability exist around it." }, { "type": "relatedTo", "target": "artifact:the-agents-are-not-the-bottleneck-you-are", "label": "The Agents Are Not the Bottleneck. You Are.", "posture": "declared-related-work", "evidence": "Owner authorization on PR #214 and verified canonical target identity in current main.", "description": "Agent execution becomes useful only when human attention, supervision, validation, and closure are managed as scarce operating resources." } ], "presentation": { "legacyChrome": { "headers": [], "footers": [] }, "evidenceSections": [ { "tag": "section", "id": "references", "role": "public-sources", "generatedPosture": "article-essential" } ], "legacyControls": [], "theme": { "capability": "explicit-light-dark", "default": "system", "rootAttribute": "data-theme", "modes": [ "system", "light", "dark" ], "generatedControl": "none", "preferenceScope": "shareplane-theme-capable-artifacts" } }, "runtime": { "javascript": "artifact-local", "siteWideJavaScript": false, "progressiveEnhancement": true, "networkAccess": false, "persistence": [ "stop-prompting-agents-theme" ] }, "validation": { "contract": "validation-contract:1.0.0", "runtimeJavaScriptException": true, "stripRuntimeScriptsAtBuild": false, "artifactAssertions": [ { "kind": "sha256", "value": "dd88c05a04ff1098a55379b61480b591ab2d98182dbda62062c12211d4c70f3b" }, { "kind": "contains", "value": "Stop Prompting Agents. Start Managing Workers." }, { "kind": "count", "token": "