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 mgallet92i/waterfall/plugin install waterfallWrote 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/mgallet92i/waterfall/wf-pm-light)<a href="https://agentmods.dev/skills/mgallet92i/waterfall/wf-pm-light"><img src="https://agentmods.dev/badge/skills/mgallet92i/waterfall/wf-pm-light/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/mgallet92i/waterfall/wf-pm-light"><img src="https://agentmods.dev/badge/skills/mgallet92i/waterfall/wf-pm-light.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 | $0.00041 | $0.02638 |
| Opus 5 | $0.00020 | $0.01319 |
| Sonnet 5 | $0.00008 | $0.00528 |
| Haiku 4.5 | $0.00004 | $0.00264 |
Grade A, and why
wf-pm-light 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 10d 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 — 232 lines — stays where its author put it; the contents beside it link to each section on GitHub.
wf-pm-light — PM solo mode subagent-light
⚠ CONSTITUTION — Règles universelles Waterfall
Lire obligatoirement avant toute action :
agents/_shared/constitution.mdInvariants universels, format SendMessage, protocole ACK, prohibitions universelles, mapping artefacts → owners, Session INV, Bash write prohibition.
Rôle
PM est le main agent en mode subagent-light. Il ne spawne pas PO/RV/QA/DS/OR.
Il pilote directement la state machine et délègue à TL-light via deux appels Agent.
Exactement 3 interactions HO (INV-001) en mode HO standard :
- Interaction n°1 : checkpoint-specs (Phase C)
- Interaction n°2 : checkpoint-tasks (Phase E)
- Interaction n°3 : validation-finale (Phase G)
L'élicitation (Phase A) précède ce comptage — elle peut être multi-questions mais ne consomme pas un round.
Combo dark_factory=on (OBS-002)
Quand .config.dark_factory == "on", dark mode l'emporte sur les checkpoints : aucune
interaction HO, PM-light auto-approuve les 3 rounds (specs / tasks / validation finale) en
tant que décideur de dernière instance. Le workflow et les artefacts restent identiques
(specs.md, design.md, tasks.md, 2 passes TL), seules les AskUserQuestion des Phases C, E, G
sont supprimées et remplacées par une décision PM loguée dans or.log.
L'élicitation (Phase A) est conservée en dark — PM a besoin d'un grill minimal pour rédiger specs.md sans ambiguïté.
Micro-tweaks PM : PM peut corriger < 5 lignes dans les fichiers implémentés, sans logique métier, sans respawn TL (INV-003 tolérance light).
PM seul auteur de specs.md. wf-auth.sh autorise PM→specs.md en mode subagent-light (EX-014).
Session INV — Premier usage de wf-orchestrate.sh
Avant tout --query ou --complete, exécuter :
bash ${CLAUDE_PLUGIN_ROOT}/scripts/wf-orchestrate.sh --help
Workflow
Boucle standard d'avancement (référencée par toutes les phases)
Les steps, leur ordre, leurs params et les auto-skips sont propriété du script — ne jamais les recopier ni les supposer (ARCH-06). Pour faire avancer la machine :
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.
- 10d ago First seen · 232 lines · 41 tokens per session scan A 9e9baf855762
wf-pm-light is a skill published in the GitHub repository mgallet92i/waterfall (2 stars, last pushed 1mo ago), licensed MIT. It adds 41 tokens to every session and 2,638 once invoked, about $0.0002 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
analyze
Use this skill after Phase 1 Plan to create or intake requirements for a task. Produces requirements with stable IDs (REQ-001, REQ-002, ...), validates each requirement against the project's scope statement, and — critically — halts and asks for UX designs and brand guidelines whenever the work touches a frontend or…
gate-signoff
Use this skill to capture a human phase-gate sign-off through chat rather than requiring the user to manually edit the gate markdown file. Prompts the user with the artifact path, collects a work-item URL (REQ / ticket / CR) as the non-trivial acknowledgment, validates the URL shape against the ticket system…
ship-it
Worktree-side implementation orchestrator for an OpenSpec change. Idempotent: gates automated scenarios on filesystem reality, owns the red-test fix loop, runs the docker harness with always-teardown, then drives ship-change inline. Escape hatch writes SHIPITBLOCKED.md. Runnable headless. Triggers: "ship it", "build…
project-init
Scaffold an unconfigured directory into a configured pi project. Interactive, profile-driven: previews the planned writes, then writes AGENTS.md, .pi/settings.json and prompt files — optionally also a knowledge base, an openspec/ scaffold, and user-global /.pi/agent/settings.json. Use on a bare directory, or when the…
execute-round
Use when a planned round handoff exists and needs to be implemented end-to-end via the 4-commit chain (entry-admin → BA design → dev body with inline CR + SE → state refresh + self-audit). Dispatches ba-designer / developer / cr-reviewer / se-contract sequentially.
debt-ops-add
Register a deferred decision in the tech-debt registry. Trigger by judgment, not a marker scan, whenever a future reader would ask "why this way?": an unmade decision, stub, loosened type, bypassed check, swallowed error, a default picked "for now", or a TODO/FIXME/HACK/XXX marker. Trigger immediately whenever you…