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.
npx skills add NVZver/claude-marketplace --skill observegit 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/skills/nvzver/claude-marketplace/observe)<a href="https://agentmods.dev/skills/nvzver/claude-marketplace/observe"><img src="https://agentmods.dev/badge/skills/nvzver/claude-marketplace/observe/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/nvzver/claude-marketplace/observe"><img src="https://agentmods.dev/badge/skills/nvzver/claude-marketplace/observe.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 | $0.00131 | $0.02626 |
| Opus 5 | $0.00066 | $0.01313 |
| Sonnet 5 | $0.00026 | $0.00525 |
| Haiku 4.5 | $0.00013 | $0.00263 |
Grade A, and why
observe 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 11d 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 — 96 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Trace. On load, print first:
=============== [observer/skills/observe/SKILL.md] [observer] ===============
Observe
Orchestrator Actor. Rides the substrate's self-paced /loop (omit the interval to self-pace) and, on each wake, reads the file changes since the previous cycle and emits feedback through the active role. All role behavior — lens, voice, cadence, difficulty rules — is data read from ../../knowledge/roles.md; this Actor applies whichever bundle is active generically and contains no per-role branch logic. It rides the existing /loop; it does not build a scheduler (../../../.lsa/VISION.md principle 9, Substrate-native first).
Goal
From a confirmed role, run a live observe-and-coach session over the user's working changes — emitting feedback that matches the active role's lens/voice/cadence each /loop cycle — until a stop trigger ends the loop with a stated reason.
Input
- The working context (current directory, open/recent files, language). Used to infer a candidate role.
- An optional role named by the user (rubber-duck, pair-programmer, interviewer, custom).
- For a custom role: a one-line lens/voice description.
- For an interviewer role: a target language and topic for the exercise.
- An optional inactivity-limit override (default: 2 consecutive no-change
/loopcycles — see Step 10). - The role bundles (lens / voice / cadence / difficulty data) from
../../knowledge/roles.md.
Steps
-
Kickoff — read the context and check for a named role (F1.2). Read the working context (current directory, open/recent files, language, test state). If the user named a role, adopt it without proposing an alternative (F1.2) and continue at Step 4. Observable result: either an adopted role, or the read context handed to Step 2.
-
Kickoff — infer one candidate role from the signal→role map (F1.1). Match the working context against this table, top-down; take the first row whose signal holds:
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.
- 11d ago First seen · 96 lines · 131 tokens per session scan A 5c97ec47825d
observe is a skill published in the GitHub repository NVZver/claude-marketplace (1 stars, last pushed 16d ago), licensed MIT. It adds 131 tokens to every session and 2,626 once invoked, about $0.0007 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
high-output-management
Manage for output using Grove's "High Output Management": a manager's output is their organization's output, raised by high-leverage activities. Use when the user mentions "high output management", "managerial leverage", "one-on-ones", "1:1 agenda", "OKRs", "performance review", "task-relevant maturity", "delegation"…
37signals-way
Build lean, opinionated products using the 37signals philosophy from "Getting Real", "Rework", and "Shape Up". Use when the user mentions "Getting Real", "Rework", "Shape Up", "37signals", "Basecamp method", "six-week cycles", "fixed time variable scope", "appetite vs estimates", "betting table", "breadboarding", "fat…
drive-motivation
Design motivation systems using Autonomy, Mastery, and Purpose (AMP) for products and teams. Use when the user mentions "intrinsic motivation", "gamification isnt working", "rewards arent working", "autonomy", "mastery", "purpose-driven", "my team is disengaged", or "how do I motivate people". Also trigger when…
weekly-briefing
This skill should be used when the user asks "what should I do this week", "weekly briefing", "run my week", "what's on my plate this week", "walk me through this week", or "give me my weekly checklist". Conducts one in-season briefing by running start/sit, waiver, and cut analysis in order. Not for a standalone…
successor-handoff
Use when orchestrating long-running autonomous work — overnight runs, multi-hour research, multi-track experiments, 10+ hour jobs — where a single Claude context window cannot hold the full run. Establishes a lean parent orchestrator, file-first subagents, and successor-handoff (fresh subagent spawned mid-run when…
style-profile
Guide a user to a personal writing-style profile - a saved selection across the catalog's Voice, Tone, Style, and Format axes - and prove it sounds like them with a generated sample before saving. Offers four intake modes (infer from your writing, recognize from examples, interview, or fill a template), confirms with…