Getting it into your agent
This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.
/plugin marketplace add geledek/enterprise-ai-transformation-skills/plugin install enterprise-ai-transformation-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/geledek/enterprise-ai-transformation-skills/process-productionization)<a href="https://agentmods.dev/skills/geledek/enterprise-ai-transformation-skills/process-productionization"><img src="https://agentmods.dev/badge/skills/geledek/enterprise-ai-transformation-skills/process-productionization/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/geledek/enterprise-ai-transformation-skills/process-productionization"><img src="https://agentmods.dev/badge/skills/geledek/enterprise-ai-transformation-skills/process-productionization.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00126 | $0.02057 |
| Opus 5 | $0.00063 | $0.01028 |
| Sonnet 5 | $0.00025 | $0.00411 |
| Haiku 4.5 | $0.00013 | $0.00206 |
Grade A, and why
process-productionization scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 111 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Process — Pilot-to-Production Playbook
Take a working prototype to production-grade. Stanford's 51-deployment study found most failures happen after the demo: missing SLOs, no fallback path, no on-call, project team disbands before product team forms. NIST RMF MANAGE 2.4 mandates off-path handling. This playbook closes the gap with 5 stages and a 15-item go-live gate.
Verdict vocabulary (stable output contract): GO / CONDITIONAL GO / NO-GO, with remediation list and week-by-week rollout schedule.
Stage 1: SLO & Eval Definition
Demos optimize for the happy path. Production needs measured floors and ceilings. Consult pilot-discipline-ng.md: every pilot needs pre-declared success metrics or it cannot graduate.
- What is the p50 and p95 latency budget? (User-perceived; include retrieval, reasoning, tool calls, render.)
- What is the minimum acceptable accuracy on the golden set? (Floor below which you roll back; size the golden set ≥200 labeled cases.)
- What is the maximum tolerable hallucination / fabrication rate? (Ceiling per 100 calls; measured on adversarial-set ≥100 cases.)
- What is the regression eval cadence? (Run on every prompt change, model version bump, retrieval-index refresh.)
- Who owns the eval suite as code? (Named individual; suite lives in CI, not a notebook.)
Output: P50_LATENCY | P95_LATENCY | ACCURACY_FLOOR | HALLUCINATION_CEILING | EVAL_OWNER
Stage 2: Fallback & Failure Design
The r10 complaint — "AI still makes mistakes, benefit unclear" — is a fallback-design failure, not a model failure. Consult imda-4-dimensions-agentic.md: structural controls (kill-switch, human-confirm, scope-fence) belong here. Consult nist-rmf-functions.md: MANAGE 2.4 requires off-path procedures for incidents.
- What is the silent-failure mode? (When the model is confidently wrong — who catches it, what signal triggers? Confidence score alone is insufficient.)
- What is the human-in-loop trigger? (Defined thresholds: confidence < X, novel input class, regulated decision, monetary value > Y.)
- What is the kill-switch latency? (Time from incident detection to system disable; target <5 minutes for high-stakes.)
- What is the deterministic fallback path? (Rule-based or human queue when AI declines; never a blank screen.)
- How are incidents logged for post-mortem? (Trace ID, input, output, ground truth, decision; retained per regulatory requirement.)
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 12d ago First seen · 111 lines · 126 tokens per session scan A 653cf0cc9d95
process-productionization is a skill published in the GitHub repository geledek/enterprise-ai-transformation-skills (9 stars, last pushed 2mo ago), licensed MIT. It adds 126 tokens to every session and 2,057 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
ai-interaction-pattern-advisor
Advises WHICH AI interaction pattern fits a product, service, or workflow — at what level of automation, with what trust mechanics and what organizational change. Acts as a senior Service Designer and AI Transformation specialist. Use whenever the user is deciding how AI should show up in an experience, or is…
ai-act-classification
Classify AI systems and GPAI models under the current EU AI Act. Use for Article 5 prohibited-practice screening, Article 6 and Annex I/III high-risk analysis, Article 6(3) exceptions, Article 50 transparency duties, GPAI duties, role allocation, deadlines, or an evidence-backed classification memo.
ai-act-compliance
Assess and build an EU AI Act compliance programme. Use for inventories, role maps, deadline readiness, Article 4 AI literacy, high-risk provider or deployer controls, GPAI governance, policy roadmaps, remediation plans, or executive compliance status.
ai-vendor-assessment
Assess an AI vendor, model provider, or supplier relationship. Use for EU AI Act role and evidence checks, GDPR processor terms, security and resilience review, model-change controls, incident routes, audit rights, contractual redlines, or a go/no-go procurement recommendation.
dpia-ai
Run or review a GDPR data protection impact assessment for an AI system. Use for Article 35 trigger screening, necessity and proportionality, profiling and automated decisions, bias and explainability risks, mitigations, DPO consultation, Article 36 prior consultation, or DPIA and EU AI Act FRIA coordination.
governance-documentation
Create, review, or organise EU AI Act governance evidence. Use for Article 11 and Annex IV technical documentation, logs, quality management, conformity records, declarations, registration, deployer records, post-market monitoring, incident files, evidence indexes, or review packs.