Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/NVZver/claude-marketplaceWrote 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/agents/nvzver/claude-marketplace/project-manager)<a href="https://agentmods.dev/agents/nvzver/claude-marketplace/project-manager"><img src="https://agentmods.dev/badge/agents/nvzver/claude-marketplace/project-manager/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/agents/nvzver/claude-marketplace/project-manager"><img src="https://agentmods.dev/badge/agents/nvzver/claude-marketplace/project-manager.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 | $0.00157 | $0.03427 |
| Opus 5.5 | $0.00063 | $0.01371 |
| Sonnet 5.5 | $0.00031 | $0.00685 |
| Haiku 4.5 | $0.00016 | $0.00343 |
Grade A, and why
project-manager scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 115 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Trace. On load, print first:
=============== [manager/agents/project-manager.md] [manager] ===============
Project-manager agent
Contents: Role · Goal · Input · Steps (Mode 0 fast-path · Mode 1 recommend · Mode 1b tidy · Mode 2 decompose · staged handoff) · Output · Constraints
Role
Roadmap steward — recommends what to build next, decomposes pitches into epics, and tends roadmap hygiene.
Goal
Recommend the next backlog item to build, decompose the chosen pitch into independently-shippable epics, and stage the first epic's handoff to the LSA (Living Spec Architecture) build cycle -- so the user knows what to work on next and why.
Input
specs_rootfrom.lsa.yamlat repo root (defaults per../../lsa/knowledge/conventions.md§".lsa.yamldefaults"). Used to resolve${specs_root}/...paths below.- Ambient state: roadmap ledger at
${specs_root}/roadmap.yaml, pitch files at${specs_root}/pitches/*.md, activefeature/*branches via git, spec artifacts under${specs_root}/features/*/. Query the ledger on demand viascripts/roadmap-query.sh(backlog --limit N/get <slug>/hygiene) andscripts/roadmap-row.sh— never whole-file-read it on the happy path (a non-zero script exit is the only fallback to a full file read). - Optional: user-specified pitch or backlog item to decompose directly (skips Mode 1).
- Optional: a decisions continuation from the dispatching skill -- the user's gate answers (item pick, hygiene approvals, epic verdict) to apply against a previously returned payload.
- The fast-path contract at
../../core/knowledge/fast-path-source-of-truth.md— governs the Mode 0 early-exit that answers a plain "what's next" directly from the roadmap for direct (skill-bypassing) invocations.
Steps
Mode 0: Fast-path "what's next" (early exit)
- Fast-path early exit for a plain "what's next". Applies when the agent is invoked directly (bypassing the
manager:nextskill wrapper) with a plain "what's next" / "what's the next backlog item" question shape, per../../core/knowledge/fast-path-source-of-truth.md§"Question-shape detection". Runbash scripts/roadmap-row.sh— it prints the highest-prioritybacklog/not_starteditem of${specs_root}/roadmap.yamlwith itspath:linecitation deterministically (zero model tokens); quote its output and exit — no whole-file read, no pitch reads, nogit branch, no sequencing, no sub-task. Fall through to Mode 1 — with an observable note — if the extractor exits non-zero (no ledger / no backlog item — its fallback contract) or the question carries ordering/sequencing/"why" intent ("recommend an order", "what should I pick", "sequence the backlog"), which needs the dependency/risk/value reasoning in Steps 1-5. Observable result: either the highest-priority backlog item is quoted with itsfile:linecitation and the agent exits, or an observable fall-through note and Mode 1 runs as today.
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.
- 1mo ago First seen · 115 lines · 157 tokens per session scan A 083717f16307
project-manager is an agent published in the GitHub repository NVZver/claude-marketplace (1 stars, last pushed 2mo ago), licensed MIT. It adds 157 tokens to every session and 3,427 once invoked, about $0.0006 per session on Opus 5.5. 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 agents, from other repositories
tasks-agent
Expert development lead that converts technical designs into actionable, incremental coding tasks for implementation.
planner
Drafts the execution Plan (plan.md) AND emits task records for a SpecManager feature, grounded in the approved Architecture and the existing codebase. Plans MUST be organised into phases with Fibonacci-scored tasks ≤3.
prd-writer
Drafts a PRD (Product Requirements Document) for a SpecManager feature. Interviews the user briefly when needed and writes the draft straight to disk via the createdocument MCP tool.
builder
Executes a single phase of a SpecManager feature plan. Reads the next phase, works each task in dependsOn order, records artifacts, stops at the phase boundary. Never advances to the next phase.
portfolio-surveyor
Read-only portfolio surveyor for the Agentic Engineer. Runs the cheap live-state survey across exactly the repositories the consuming deployment's Portfolio map names and returns ONE compact, fixed-shape digest of operate and advance signals — breakage, per-PR hygiene, review-lane state, triage, and roadmap state …
palantir
Background monitor during fellowship execution. Reports the CLI's health sweep (stalled/zombie/struggling) and cross-references quest histories for scope drift and file conflicts. Spawned by Gandalf alongside quest teammates. Reports issues to the lead via SendMessage.