asdc163

@asdc163 source repo

79 skills
0 followers
17 hours ago last updated

▌
Pm Proof To Share · asdc163 bundle
Turn a verified product or skill release into an evidence-backed, channel-aware English share pack with a clear user job, proof ledger, try path, boundaries, feedback ask, and learning writeback. Use when a maintainer needs public discovery material without inventing adoption or growth claims.

0
installs
▌
Pm Source To Test · asdc163 bundle
Turn raw product feedback, research notes, support tickets, or product observations into a source-linked PM review with explicit claims, limitations, and one smallest test. Use when a PM needs to challenge a conclusion, choose what to test next, or hand off a decision without inventing evidence.

0
installs
▌
Pm Feedback To Fix · asdc163 bundle
Turn a de-identified product observation into a bounded reproduction path, smallest fix or experiment, acceptance checks, release and rollback notes, and a learning writeback. Use when a PM or maintainer needs to move from session feedback to an evidence-safe action.

0
installs
▌
Pm AI Task Boundary · asdc163 bundle
Decide how an AI capability should divide work between a person and an AI system by mapping the user job to a SCAN zone, autonomy level, permissions, approval points, fallback, evaluation slices, and a smallest safe pilot. Use before an AI assistant, copilot, agent, automation, RAG, or AI workflow moves into product scope or implementation.

0
installs
▌
Pm Decision To Spec · asdc163 bundle
Turn an evidence-backed product decision into a bounded Product Decision Packet with scope, UX states, acceptance criteria, measurement, rollout, and implementation handoff. Use when a PM needs to move from a decision or readout to a reviewable build slice without inventing evidence.

0
installs
▌
Pm Release To Learn · asdc163 bundle
Turn a verified product or skill release into a bounded rollout-and-learning plan with audience, exposure, observation window, success and guardrail signals, rollback trigger, feedback capture, and a next decision. Use after release proof exists and before or during a pilot when a PM needs to learn without turning traffic or stars into adoption claims.

0
installs
▌
Pm AI Model To Route · asdc163 bundle
Turn an AI model, provider, or version choice into a source-bounded route contract covering user-job slices, candidate capabilities, manual or automatic selection, eligibility, quality, safety, privacy, cost, latency, reliability, quota, fallback, route receipts, version drift, rollback, and a Ship, Pilot, Iterate, Hold, Rollback, or Need evidence decision. Use when a PM reviews model selection, multi-model routing, provider changes, model aliases, fallback paths, or an AI route change before or after release.

0
installs
▌
Pm AI Output To Eval · asdc163 bundle
Use when a model output needs a repeatable product-quality decision beyond schema validity. Define the evaluation unit, source/reference, deterministic and human/model oracles, negative slices, abstention, disagreement, drift, denominator, release gate, rollback, and platform migration without treating one score as truth.

0
installs
▌
Pm Outcome To Metric · asdc163 bundle
Turn a product outcome or AI product goal into an evidence-bounded metric contract with a primary measure, denominator, window, guardrails, instrumentation gaps, and a decision rule. Use before an experiment, rollout, or evaluation when a PM needs to define what success means without inventing baselines or analytics evidence.

0
installs
▌
Pm Trend To Decision · asdc163 bundle
Turn a dated AI, platform, developer-tool, or market change note into a source-linked PM decision brief with impact, uncertainty, and one smallest validation. Use when a PM needs to decide whether a change matters, who it affects, or what to test next.

0
installs
▌
Pm AI Data To Purpose · asdc163 bundle
Turn an AI feature or agent data flow into a source-bounded data-purpose and lifecycle contract covering collection, use stages, provenance, authority, sensitivity, minimization, tenant scope, retention, deletion, correction, reuse, evaluation, training, third-party egress, verification, rollback, and a Ship, Iterate, Hold, Rollback, or Need evidence decision. Use when a PM reviews what AI data may be collected, shown to a model, logged, exported, reused, or deleted, especially across providers, tools, tenants, evaluation sets, or material product changes.

0
installs
▌
Pm AI Evaluation Plan · asdc163 bundle
Turn an AI feature goal and available evidence into a bounded evaluation plan with test slices, rubric, failure taxonomy, judge boundary, guardrails, fallback, and release gate. Use when a PM needs to decide what to measure before building, comparing, or promoting an AI capability.

0
installs
▌
Pm AI Risk To Control · asdc163 bundle
Turn an AI product or agent hazard into a reviewable risk-to-control contract with affected users and assets, harm paths, preventive/detective/corrective controls, verification oracles, residual risk, ownership, fallback, rollback triggers, and a Ship, Pilot, Hold, Rollback, or Need evidence decision. Use for pre-launch or pre-change risk reviews; do not treat a risk register as safety, security, legal, compliance, or adoption evidence.

0
installs
▌
Pm Opportunity To Bet · asdc163 bundle
Turn multiple evidence-backed opportunity candidates into one bounded product bet with a source ledger, user job, assumptions, opportunity cost, smallest validation, non-goals, and a stop or revise rule. Use when a PM must choose what to pursue next without turning an unsourced score, single signal, or market guess into a roadmap commitment.

0
installs
▌
Pm AI Approval To Flow · asdc163 bundle
Turn an AI or agent action proposal into an evidence-bounded approval flow with risk classification, preview and diff, permission boundary, approve/reject/edit/defer states, durable receipt, recovery path, and evaluation or release gate. Use before an AI-assisted workflow can send, publish, delete, change access, spend money, or create another consequential side effect.

0
installs
▌
Pm AI Feedback To Eval · asdc163 bundle
Turn an AI user correction, preference, thumbs-down report, escalation, or reviewed trace into a privacy-safe, evidence-bounded evaluation case with provenance, observation and label separation, oracle, slice, calibration, contamination checks, dataset destination, fallback, and release decision. Use when a PM needs to decide whether real feedback should become a golden, regression, negative-routing, red-team, canary, holdout, product-fix, or no-action artifact.

0
installs
▌
Pm AI Memory To Policy · asdc163 bundle
Turn a proposed AI or agent memory feature into a source-bounded memory policy for user value, write and read eligibility, provenance, scope, freshness, privacy, retention, correction, deletion, export, poisoning defense, evaluation, fallback, and a Ship, Iterate, Hold, Rollback, or Need evidence decision. Use when a product may remember user, project, tenant, or agent facts across sessions.

0
installs
▌
Pm AI Output To Schema · asdc163 bundle
Use when an AI response or function-call argument must cross a schema boundary. Define route, provider/model/SDK, schema and version, required evidence, refusal/incomplete/parse/drift states, bounded recovery, user-visible fallback, authority separation, evaluation, and release evidence without treating valid JSON as truth or authorization.

0
installs
▌
Pm AI Skill To Package · asdc163 bundle
Use when an AI capability may be packaged as a reusable agent skill. Produce a source-bounded package contract covering identity, discovery triggers, progressive disclosure, resources and scripts, permissions, surface compatibility, provenance, versioning, verification, disablement, rollback, and a truthful release decision.

0
installs
▌
Pm AI Task To Progress · asdc163 bundle
Use when an AI or agent task may run across waits, retries, approvals, process restarts, or host changes. Produce a source-bounded task-to-progress contract with stable identity, honest progress evidence, pause/resume/cancel/retry controls, terminal-state proof, recovery, privacy, and a manual fallback.

0
installs
▌
Pm AI Tool To Contract · asdc163 bundle
Turn an AI or agent tool or MCP integration into a source-bounded agent-facing contract for the user job, purpose, scope, namespacing, input schema, examples, output signal-to-noise, permissions, side effects, errors, idempotency, provenance, prompt-injection handling, evaluation slices, rollout, and a ship, hold, or rollback decision. Use when a PM reviews a new tool, tool change, connector, function, or agent workflow.

0
installs
▌
Pm AI Claim To Citation · asdc163 bundle
Turn an AI-generated answer, research brief, or agent output into a source-bounded claim-to-citation contract covering claim segmentation, entailment, citation coverage and placement, source authority and freshness, conflict, uncertainty, privacy, prompt-injection boundaries, reader verification, abstention, evaluation, fallback, and release decision. Use when a PM needs to decide whether an AI output is supportable, must be qualified, or must not be shown.

0
installs
▌
Pm AI Cost To Guardrail · asdc163 bundle
Turn an AI or agent cost or latency signal into a source-bounded cost ledger, successful-outcome denominator, p50 and p95 latency budget, quality and trust guardrails, routing or scope options, and a ship, hold, or rollback decision. Use when a PM evaluates model, prompt, retrieval, tool, agent, context, caching, batching, or fallback changes against a real product budget.

0
installs
▌
Pm AI Policy To Product · asdc163 bundle
Turn an external AI policy, standard, regulation, contract, or governance requirement into a source-bounded product control map with applicability, actors, product surfaces, owners, evidence receipts, exceptions, freshness, and a truthful release route. Use when a PM must translate policy language into product work without giving legal advice or claiming compliance.

0
installs
▌
Pm AI Program To Result · asdc163 bundle
Use when an AI workflow may let a model-generated program call eligible tools and a product manager must define the direct-versus-programmatic boundary, parent/program/child caller linkage, allowed tools, input and output schemas, budgets, pause and continuation, final-message completeness, citations, recovery, and the boundary between program output and a verified user outcome.

0
installs
▌
Pm AI Prompt To Version · asdc163 bundle
Use when a prompt change may alter a user-facing AI or agent workflow. Produce a source-bounded prompt version contract covering identity, input and output contracts, change diff, baseline and candidate evidence, rollout, cost and latency guardrails, data boundaries, rollback, and a truthful release decision.

0
installs
▌
Pm AI Workflow To Scale · asdc163 bundle
Turn an AI workflow demo, validation, or pilot into an evidence-bounded Explore, Validate, Pilot, Scale, Narrow, Hold, or Retire decision using accepted outcomes, guardrails, reliability, cost, demand, capacity, ownership, rollout, and rollback.

0
installs
▌
Pm Interview To Insight · asdc163 bundle
Turn de-identified interview notes, usability sessions, or observed workflows into an evidence-bounded insight map, contradiction log, and one next learning question or smallest validation. Use when a PM needs to learn from conversations without mistaking a quote, preference, or single session for a segment-wide finding.

0
installs
▌
Pm AI Drift To Diagnosis · asdc163 bundle
Use when an AI product's quality, behavior, cost, latency, coverage, or completion signal changes across time and the PM must distinguish real drift from input mix, source, oracle, product, model, policy, instrumentation, or operational explanations before choosing an eval, intervention, hold, or rollback route.

0
installs
▌
Pm AI Review To Capacity · asdc163 bundle
Turn one AI output or agent job into a source-bounded human-review operations packet with a review unit, coverage or sampling policy, queue demand, reviewer capacity, quality, privacy, escalation, fallback, economics, owners, and a Cover required, Sample and monitor, Add capacity, or Hold route without claiming safety, quality, adoption, or production readiness.

0
installs
▌
Pm AI Value To Retention · asdc163 bundle
Use when a PM needs to test whether an AI product creates meaningful repeat value and decide how to handle retention, reactivation, suppression, and trust without optimizing notification clicks.

0
installs
▌
Pm AI Vendor To Decision · asdc163 bundle
Turn one external AI provider or vendor dependency into a source-bounded PM decision packet for a named user job and scope. Keep provider identity, service/contract role, data use and region, model/tool lifecycle, security/privacy evidence, availability/support/limits, cost, lock-in/portability, ownership, implementation receipts, and exit conditions visible without ranking vendors, giving legal advice, or claiming a recommendation or production readiness.

0
installs
▌
Pm AI Vision To Decision · asdc163 bundle
Use when an AI product reads images, PDFs, charts, screenshots, scans, or other visual artifacts and the team must define what was received, what was actually extracted, where each claim came from, what is ambiguous or illegible, how people can verify it, and whether the route is safe to pilot. Turn multimodal input into a source-bounded decision contract without treating fluent descriptions as visual accuracy.

0
installs
▌
Pm Experiment To Readout · asdc163 bundle
Turn a bounded product test result into an evidence-aware PM readout with metric, guardrail, decision rule, limitations, and one next action. Use when a PM needs to decide whether to continue, change, stop, or hold after a prototype, experiment, pilot, or evaluation.

0
installs
▌
Pm AI Code Run To Sandbox · asdc163 bundle
Use when an AI feature may generate, inspect, modify, or execute code and a product manager needs an explicit sandbox, filesystem, network, package, secret, approval, cancellation, artifact, and verification contract before a run is allowed.

0
installs
▌
Pm AI Context To Contract · asdc163 bundle
Turn an AI or agent context change into a source-bounded context contract covering instructions, knowledge, tools, memory, state, query, selection rules, freshness, privacy, token budget, compaction, evaluation slices, fallback, and a ship, hold, or rollback decision. Use when a PM reviews prompts, retrieval, tool schemas, MCP, conversation history, memory, context assembly, or long-running agent behavior.

0
installs
▌
Pm AI Handoff To Recovery · asdc163 bundle
Turn an AI assistant or agent's uncertainty, missing authority, blocked action, tool failure, or high-impact boundary into a privacy-safe human or specialist handoff and recovery contract with a minimal context packet, destination, owner, permissions, visible states, acknowledgement, resume rule, rollback, and learning writeback. Use for pre-launch or pre-change design of escalation and continuation; do not confuse a handoff with approval, identity, incident response, resolution, or adoption evidence.

0
installs
▌
Pm AI Incident To Runbook · asdc163 bundle
Turn an AI or agent incident signal into a critical-journey impact map, evidence-bounded severity, safe containment, recovery runbook, communication boundary, verification and reopen gate, and learning writeback. Use when several runs, users, or operational signals suggest a journey-level failure and a PM needs an actionable response without inventing prevalence, root cause, or production readiness.

0
installs
▌
Pm AI Intent To Discovery · asdc163 bundle
Use when a PM needs to decide when an AI capability should be surfaced, selected, declined, or handed to a manual route based on user intent, context, host capability, and permission.

0
installs
▌
Pm AI Output To Interface · asdc163 bundle
Turn an AI or agent result into a source-bounded output-to-interface contract that chooses text, structured data, a declarative UI, or an action proposal; maps data to trusted components, states, fallback, provenance, permissions, accessibility, evaluation, and release decision. Use when a PM is designing generative UI, structured-output views, MCP Apps, ChatGPT Apps SDK widgets, agent result cards/forms/dashboards, or adaptive AI workflows before engineering.

0
installs
▌
Pm AI Provenance To Trust · asdc163 bundle
Use when an AI-generated or AI-edited asset needs an origin, history, integrity, signer, watermark, or Content Credentials decision that users can understand. Produce a bounded provenance and trust contract with asset identity, bindings, verification states, transformation gaps, privacy, downstream decision separation, recovery, and a ship, pilot, hold, or rollback decision. Do not treat provenance as proof of truth, safety, authorship, identity, or legal rights.

0
installs
▌
Pm AI Realtime To Session · asdc163 bundle
Use when an AI product has live voice, realtime audio, translation, or streaming transcription and the team needs a bounded session contract. Choose the session type, identity, authority, transport, turn-taking, interruption, tools, consent, recovery, metrics, evaluation slices, and release decision without claiming a demo is a working product.

0
installs
▌
Pm AI System To Inventory · asdc163 bundle
Turn one AI capability, agent, model-backed feature, or vendor integration into a source-bounded lifecycle inventory record with identity, purpose, actor and owner, users and affected parties, surfaces, dependencies, data and sources, model and tool versions, permissions, policy, evidence, change and incident links, review cadence, and retirement. Use when a PM needs to know what AI exists and who is accountable without treating a complete record as proof of safety, compliance, adoption, value, or production readiness.

0
installs
▌
Pm AI Trace To Regression · asdc163 bundle
Turn an AI or agent failure trace, tool error, user correction, or guardrail event into an evidence-bounded failure classification, containment step, minimal reproduction, regression case, owner, and release decision. Use after an AI run, pilot, evaluation, or production observation when a PM needs to separate trace facts from root-cause hypotheses without inventing prevalence or safety claims.

0
installs
▌
Pm AI Value To Investment · asdc163 bundle
Turn one AI workflow into an evidence-bounded value-to-investment brief with a successful-work unit, full cost ledger, dependability, value assumptions, scenarios, sensitivity, capacity, and an Invest, Test, Narrow, Hold, or Stop decision.

0
installs
▌
Pm AI Workflow To Package · asdc163 bundle
Turn a tested AI workflow into an evidence-bounded operating package that another person can repeat, review, support, maintain, change, or retire without overstating adoption, value, safety, or production readiness.

0
installs
▌
Pm AI Identity To Boundary · asdc163 bundle
Turn an AI or agent actor into a source-bounded identity and authorization contract covering principals, authentication, delegation, resource scope, tenant boundaries, least privilege, approval interaction, credential and token lifecycle, revocation, attribution, audit receipts, evaluation, fallback, and a Ship, Iterate, Hold, Rollback, or Need evidence decision. Use when an agent may act for a user, service, workspace, tenant, or connector.

0
installs
▌
Pm AI Improvement To Route · asdc163 bundle
Turn an observed AI product quality, trust, cost, latency, coverage, or completion gap into a source-bounded choice of improvement route across prompt, context, retrieval, tools, model, data, UX, or fine-tuning. Use before committing to one technique; require a user job, failure localization, paired evaluation, permission boundary, owner, stop rule, and rollback.

0
installs
▌
Pm AI MCP To Authorization · asdc163 bundle
Use when an AI product connects to an MCP server or agent connector and the team needs a source-bounded authorization contract for resource, issuer, consent, scope, tool side effects, token lifecycle, task isolation, and recovery.

0
installs
▌
Pm AI Monitor To Oversight · asdc163 bundle
Turn an AI or agent monitor signal into a bounded human-oversight contract with observation scope, coverage, latency, review states, containment, control evaluations, privacy boundaries, and an honest safety-case evidence decision.

0
installs
▌
Pm AI Research To Evidence · asdc163 bundle
Use when an AI product searches the web, files, or connected data to answer a complex question and the team must define the decision, source policy, evidence ledger, uncertainty, tool boundary, review path, and release gate. Turn agentic research into a reviewable product contract without treating a long report or fluent citations as proof.

0
installs
▌
Pm AI Run To Observability · asdc163 bundle
Turn an AI or agent run into a source-bounded observability contract covering run, session, task, trace, span, and event identity; prompt, tool, approval, MCP, and network evidence; provenance, identity, scope, policy, outcome, guardrail, latency, cost, privacy, retention, diagnosis, fallback, and release decision. Use when a PM needs to make an AI or agent workflow diagnosable after deployment.

0
installs
▌
Pm AI Workflow To Adoption · asdc163 bundle
Turn a tested AI workflow into an evidence-bounded team introduction and adoption plan with a real work moment, support and fallback, enablement, evidence ownership, feedback-to-change, and a continue, revise, pause, stop, or broader-use decision.

0
installs
▌
Pm AI Workflow To Evidence · asdc163 bundle
Turn scattered notes, workflow artifacts, tests, adoption observations, metrics, and stakeholder feedback into a source-bounded AI workflow evidence packet. Separate what is measured, observed, reported, estimated, planned, inferred, or unknown; preserve contribution and conflict; and choose Capture next, Proceed bounded, Hold, or Do not claim without inventing adoption, value, causality, safety, or production readiness.

0
installs
▌
Pm AI Adoption To Diagnosis · asdc163 bundle
Diagnose why one tested AI workflow is not becoming repeatable useful work, separate observed behavior from assumed cause, and choose the smallest evidence-bounded intervention or escalation. Use when adoption is stalled or uneven across a real team or workflow; do not turn usage, attendance, or a single quote into an adoption, value, quality, safety, or causality claim.

0
installs
▌
Pm AI Content To Moderation · asdc163 bundle
Use when an AI product must turn a content policy into a reviewable moderation workflow with a bounded taxonomy, severity and action matrix, timing, human review, appeals, false-positive and false-negative slices, privacy controls, and a ship, pilot, hold, or rollback decision. It separates provider capability from product evidence and does not implement or prove a moderation classifier.

0
installs
▌
Pm AI Portfolio To Sequence · asdc163 bundle
Turn several AI workflow or capability candidates into an evidence-bounded portfolio sequence with value models, foundations, dependencies, capacity, concurrency limits, stage gates, and Start, Next, Parallel, Hold, Stop, or Retire decisions.

0
installs
▌
Pm AI Review To Calibration · asdc163 bundle
Turn human review or model-judge scoring of an AI output into a source-bounded calibration contract covering the user job, artifact versions, rubric criteria, anchor examples, blind labels, reviewer agreement, judge comparison, adjudication, bias, drift, privacy, release thresholds, fallback, and writeback. Use when a PM needs to decide whether review evidence is consistent enough to support an AI quality, rollout, or release decision.

0
installs
▌
Pm AI Tool Call To Recovery · asdc163 bundle
Use when an AI agent or host emits one or more tool calls and a product manager must define request/result correlation, argument validation, execution boundaries, parallel result handling, retry and idempotency rules, late or duplicate results, user-visible recovery, and the boundary between a tool result and a verified business outcome.

0
installs
▌
Pm AI Workflow To Readiness · asdc163 bundle
Decide whether one real AI workflow candidate is ready to test now, needs more validation, should be sequenced later, or should be avoided for now. Use the supplied job, owner, value, complexity, risk, dependencies, user and technical readiness, human boundary, support, and smallest-test evidence without inventing adoption, value, safety, or production claims.

0
installs
▌
Pm AI Change To Revalidation · asdc163 bundle
Turn a proposed change to an AI workflow, source, prompt, model, tool, policy, permission, audience, or owner into an evidence-bounded impact map, revalidation plan, release or hold decision, controlled rollout boundary, and rollback route.

0
installs
▌
Pm AI Outcome To Improvement · asdc163 bundle
Turn an AI or agent proposal, human correction, downstream artifact, or verified external outcome into an evidence-bounded improvement finding. Use when a PM must separate model error from preference, workflow noise, product support, source or mapping issues, and downstream state; require identity joins, review, grouping, denominator, owner, next eval or fix, and rollback.

0
installs
▌
Pm AI Retrieval To Grounding · asdc163 bundle
Turn a search, file-search, vector-store, RAG, or grounding proposal into a source-bounded PM contract covering source eligibility, query construction, retrieval, ranking, evidence sufficiency, citations, abstention, privacy, evaluation, fallback, and release decision. Use when a team needs to decide what an AI system may retrieve and when an answer is supportable.

0
installs
▌
Pm AI Signal To Intervention · asdc163 bundle
Turn an online AI quality, safety, trust, cost, latency, policy, or behavior signal into an evidence-bounded intervention decision with scope, owner, TTL, user communication, verification, recovery, rollback, and learning writeback. Use when a PM must decide whether a live signal means observe, investigate, qualify, narrow, gate, pause, rollback, or restore without treating one noisy event or a dashboard count as an incident or release verdict.

0
installs
▌
Pm AI Subagent To Delegation · asdc163 bundle
Use when an AI workflow may delegate a bounded subtask to a specialist agent and a product manager needs to choose the delegation route, preserve least-privilege context and authority, define ownership, validate the result, and recover safely.

0
installs
▌
Pm AI Translation To Meaning · asdc163 bundle
Use when an AI product translates live speech, text, captions, or documents across languages and the team must protect meaning, intent, terminology, uncertainty, privacy, and user outcomes. Define source/target locale, context, ambiguity, correction, evaluation slices, fallback, and release evidence without treating fluent output or one score as semantic parity.

0
installs
▌
Pm AI Computer Use To Control · asdc163 bundle
Use when an AI agent may observe or operate a graphical user interface. Produce a source-bounded computer-use control contract with observation mode, action scope, postconditions, human stop points, sensitive-screen and prompt-injection boundaries, mismatch recovery, evaluation slices, and a truthful release decision.

0
installs
▌
Pm AI First Use To Activation · asdc163 bundle
Use when a PM needs to turn an AI capability into a bounded first-use journey, meaningful value oracle, activation hypothesis, instrumentation contract, and safe learning decision.

0
installs
▌
Pm AI Guardrail To Enforcement · asdc163 bundle
Use when an AI or agent workflow needs a guardrail contract that maps each input, output, tool, handoff, approval, or runtime boundary to its timing, enforcement action, tripwire, failure state, evidence, recovery, and residual risk.

0
installs
▌
Pm AI Tool Search To Selection · asdc163 bundle
Use when an AI agent has a large or changing tool catalog and a product manager must define what is searchable, which candidates are eligible, when tools are deferred or loaded, how hosted and client-owned discovery differ, when the agent should abstain, and how selection stays separate from authorization, execution, outcome, and adoption.

0
installs
▌
Pm AI Model Change To Migration · asdc163 bundle
Use when an AI product faces a model, provider, endpoint, snapshot, lifecycle, capability, price, latency, or serving-behavior change. Produce a source-bounded migration decision with identity, impact, baseline and candidate comparison, safety and cost gates, rollout, fallback, rollback, and explicit evidence limits.

0
installs
▌
Pm AI Orchestration To Contract · asdc163 bundle
Turn a multi-step AI or agent workflow into a source-bounded orchestration contract with explicit topology, step ownership, transitions, control budgets, side-effect boundaries, failure recovery, evaluation slices, and a release decision. Use before implementing or materially changing an AI workflow when the team must decide what the model, deterministic code, tools, specialists, and people do next.

0
installs
▌
Pm AI Uncertainty To Experience · asdc163 bundle
Turn AI uncertainty, ambiguity, partial evidence, delay, conflict, or failure into a user-visible experience contract with honest progress, provenance, controls, clarification, recovery, accessibility, trust evaluation, fallback, and release evidence. Use when an AI feature needs to show what is known, unknown, blocked, or ready for a human decision without treating model confidence as truth.

0
installs
▌
Pm AI Agent Elicitation To Input · asdc163 bundle
Use when an AI agent needs a missing fact, choice, clarification, or user input during a tool or task flow. Produce a source-bounded elicitation contract for purpose, provenance, schema, sensitivity, user controls, response states, timeout, validation, recovery, and the boundary to approval or side effects.

0
installs
▌
Pm AI Recommendation To Decision · asdc163 bundle
Use when an AI product presents a recommendation, ranking, triage suggestion, or plan that a person may accept, edit, reject, defer, compare, or hand off. Produce a source-bounded recommendation-to-decision contract that keeps evidence, uncertainty, user choice, and consequential execution separate.

0
installs
▌
Pm AI Independent Eval To Release · asdc163 bundle
Turn an independent or third-party AI evaluation into an evidence-bounded release decision covering claim type, evaluator independence, system and harness configuration, budget, validity hazards, access and publication boundaries, remediation, and rollback. Use when a PM needs to commission, review, or interpret an external evaluation of an AI model, agent, capability, safeguard, or product before deployment; do not treat one report, benchmark score, red-team exercise, or provider statement as proof of safety, truth, adoption, or production readiness.

0
installs
▌
Pm AI Prompt Injection To Defense · asdc163 bundle
Turn a suspected or observed prompt injection, indirect injection, tool poisoning, or untrusted agent/MCP content path into an evidence-bounded attack-path and defense contract with trust boundaries, authority separation, least-privilege controls, negative evals, rollback, and a Ship, Pilot, Hold, Block, or Need evidence decision. Use when a PM reviews agent, RAG, MCP, retrieval, tool, browser, delegated-agent, memory, or external-content behavior that may influence an AI decision or side effect; do not treat a model refusal, scanner result, or completed checklist as proof of product security.

0
installs
▌
Pm AI Background Run To Supervision · asdc163 bundle
Use when a user delegates an AI or agent task that may continue after the current interaction, run asynchronously, resume later, or start on a schedule. Produce a source-bounded supervision contract for scope, autonomy, state, checkpoints, pause, cancellation, expiry, notification, result review, retention, budget, fallback, and recovery.

0
installs
▌
Pm AI Preference To Personalization · asdc163 bundle
Use when an AI product may adapt to a user's preference, custom instruction, memory, contextual fact, or inferred trait. Produce a source-bounded personalization control contract with type, purpose, scope, freshness, correction, deletion, opt-out, temporary use, conflict handling, evaluation, and safe fallback.

0
installs

79 published skills