Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/CES-Ltd/Luminpx agentmods add skills/ces-ltd/lumi/flow-deliverWrote 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/ces-ltd/lumi/flow-deliver)<a href="https://agentmods.dev/skills/ces-ltd/lumi/flow-deliver"><img src="https://agentmods.dev/badge/skills/ces-ltd/lumi/flow-deliver.svg" alt="Measured on agentmods" 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 | $0.00071 | $0.07722 |
| Opus 5 | $0.00036 | $0.03861 |
| Sonnet 5 | $0.00014 | $0.01544 |
| Haiku 4.5 | $0.00007 | $0.00772 |
Grade A, and why
flow-deliver scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 8d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
This is a copy
94% identical to flow-deliver — 197 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 917 lines — stays where its author put it; the contents beside it link to each section on GitHub.
This file is generated from a template. Edit the
.tmplfile, not this file directly. Runscripts/gen-skill-docs.shto regenerate after changes.
Pre-Delivery: State Check
Before starting delivery:
- Read
.octo/STATE.mdto verify Develop phase complete - Update STATE.md:
- current_phase: 4
- phase_position: "Delivery"
- status: "in_progress"
# Verify Develop phase is complete
if [[ -f ".octo/STATE.md" ]]; then
develop_status=$("${HOME}/.claude-octopus/plugin/scripts/octo-state.sh" get_phase_status 3)
if [[ "$develop_status" != "complete" ]]; then
echo "⚠️ Warning: Develop phase not marked complete. Consider completing development first."
fi
fi
# Update state for Delivery phase
"${HOME}/.claude-octopus/plugin/scripts/octo-state.sh" update_state \
--phase 4 \
--position "Delivery" \
--status "in_progress"
⚠️ EXECUTION CONTRACT (MANDATORY - CANNOT SKIP)
This skill uses ENFORCED execution mode. You MUST follow this exact sequence.
STEP 1: Detect Work Context (MANDATORY)
Analyze the user's prompt and project to determine context:
Knowledge Context Indicators:
- Document terms: "report", "presentation", "PRD", "proposal", "document", "brief"
- Quality terms: "argument", "evidence", "clarity", "completeness", "narrative"
Dev Context Indicators:
- Code terms: "code", "implementation", "API", "endpoint", "function", "module"
- Quality terms: "security", "performance", "tests", "coverage", "bugs"
Also check: What is being reviewed? Code files -> Dev, Documents -> Knowledge
Capture context_type = "Dev" or "Knowledge"
Step 1b: Detect Dev Subtype (if Dev context)
When context_type is Dev, determine the subtype to inject domain-appropriate validation criteria into the review prompt. Append the matching validation supplement after the user's prompt when calling orchestrate.sh in Step 4.
| Subtype | Trigger keywords | Validation supplement |
|---|---|---|
frontend-ui |
"page", "widget", "component", "UI", "HTML", "CSS", "form", "dashboard", "layout" | Verify: all referenced files exist (scripts, stylesheets, images). Check ARIA labels and roles, keyboard navigability, touch target sizes (44px min). Flag innerHTML usage. Confirm progressive enhancement (fallbacks for navigator.share, localStorage, etc). Test self-containment: does this work if opened/run with zero setup? |
cli-tool |
"CLI", "command-line", "terminal", "script", "flag", "argument" | Verify: --help flag works, exit codes are meaningful (0/1/2), stderr vs stdout used correctly, argument edge cases handled (missing args, invalid input, --unknown-flag). |
api-service |
"API", "endpoint", "REST", "GraphQL", "gRPC", "server", "route" | Verify: input validation at every endpoint, consistent error response format, auth on protected routes, rate limiting considered, schema/contract documented. |
infra |
"deploy", "terraform", "docker", "CI", "pipeline", "Kubernetes", "helm" | Verify: operations are idempotent, no hardcoded secrets, rollback path exists, health checks included, destroy operations require confirmation. |
data |
"ETL", "pipeline", "migration", "schema", "database", "SQL" | Verify: migrations are reversible, data validation at ingestion, backup strategy documented, no data loss on failure. |
general |
Default if no subtype matches | No supplement — use standard review criteria. |
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.
- 8d ago First seen · 917 lines · 71 tokens per session scan A 47c3c7d21ebe
flow-deliver is a skill published in the GitHub repository CES-Ltd/Lumi (31 stars, last pushed 4mo ago), licensed MIT. It adds 71 tokens to every session and 7,722 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. It is 94% identical to flow-deliver, differing in 197 lines, and is treated as a copy.
Other skills, from other repositories
pr-review
Review open PRs two ways - default is a per-PR deep review with severity-tagged findings, inline comments, and a verdict; --survey runs a risk-tiered triage digest of what's safe to merge first.
pr-triage
First-touch triage for external pull requests - verdict, label, and a welcoming comment within minutes of open.
[REPLACE: SKILL_NAME]
First-touch review of newly opened PRs on [REPLACE: WATCHEDREPO] — verdict + welcoming comment + label.
code-review-excellence
Master effective code review practices to provide constructive feedback, catch bugs early, and foster knowledge sharing while maintaining team morale. Use when reviewing pull requests, establishing review standards, or mentoring developers.
multi-reviewer-patterns
Coordinate parallel code reviews across multiple quality dimensions with finding deduplication, severity calibration, and consolidated reporting. Use this skill when organizing multi-reviewer code reviews, calibrating finding severity, or consolidating review results.
fetch-tweets
Search and curate X/Twitter behind one selector - keyword, topic roundup, a single or tracked-account digest, an X list, or the AI-agent buzz preset - clustered into signal-scored sub-narratives.