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 agentmods add skills/missingpackage/nightshift/peripheral-visionnpx skills add MissingPackage/nightshift --skill peripheral-visiongit clone --depth 1 https://github.com/MissingPackage/nightshiftWhat 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 | $0.00047 | $0.00452 |
| Opus 5 | $0.00023 | $0.00226 |
| Sonnet 5 | $0.00009 | $0.00090 |
| Haiku 4.5 | $0.00005 | $0.00045 |
Grade A, and why
peripheral-vision 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 yesterday.
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.
What it actually says
Peripheral Vision
Overview
While working you see things outside the task: rot, contradictions, tripwires. By default that signal is discarded to "stay concise". This skill un-suppresses the reporting — it cannot create noticing that didn't happen (for deep sweeps use the scout agent instead).
The scan (30 seconds, before the completion report)
Ask yourself, strictly about things you ALREADY touched or read this session:
- Contradictions: did any doc/config/comment contradict observed reality? (stale branch names, wrong README claims, dead env vars)
- Tripwires: did you step around something that will bite the next change? (half-applied pattern, duplicated constant, fragile ordering, TODO-load-bearing)
- Rot: files/deps/services referenced but absent, or present but unused?
- Blast radius: does your change silently affect a consumer you didn't touch?
Reporting rules (the shape)
Append to the completion report:
**Noticed (outside task):**
1. <observation> — <why it's load-bearing> — <suggested owner: docket/Linear/now>
- Max 3 items. More than 3 means you're listing, not judging.
- Each item must pass: "would the user act differently knowing this?" If not, drop it.
- Never silently fix out-of-scope findings; report + log (project docket or Linear).
- Nothing qualified → write nothing. An empty ritual line trains the reader to skip the section.
Common mistakes
- Padding with style nits to fill the quota — the cap is a maximum, not a target.
- Re-reporting the same known issue every session — check the docket first; repeats reference the existing entry.
- Turning the scan into a new exploration — this reports what you already saw; it spends zero new tool calls (that's the scout agent's job).
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.
- yesterday First seen · 40 lines · 47 tokens per session scan A 5cbfdf52bf49
peripheral-vision is a skill published in the GitHub repository MissingPackage/nightshift (2 stars, last pushed 16d ago), licensed MIT. It adds 47 tokens to every session and 452 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
baton-setup
One-time setup and health check for fable-baton. Use when the user asks to set up, configure, verify, or troubleshoot fable-baton - sets the default model to "best" (Fable 5 with Opus fallback) in /.claude/settings.json and verifies the plugin is fully installed.
licensing-tiers-data-governance
Implement subscription tiers with field-level access control, feature gating, rate limiting, and compliance tracking. Design data governance systems that enforce different access levels, retention policies, and regulatory requirements based on user subscription tier (Free, Pro, Enterprise).
data-infrastructure-at-scale
Build data infrastructure that scales from prototype to production. Use when architecting data pipelines, choosing data stores, planning for high throughput, or migrating to distributed systems. Covers caching, replication, sharding, message queues, and data lake architecture.
data-provenance
Track data lineage and provenance from source to consumption. Use when auditing data flows, debugging data quality issues, ensuring compliance (GDPR, SOX), or understanding data dependencies. Covers lineage tracking, impact analysis, data catalogs, and metadata management.
multi-source-data-conflation
Merge and reconcile data from multiple sources into a unified view. Use when integrating APIs, consolidating databases, building data warehouses, or creating master data. Covers entity resolution, conflict resolution, data quality, and real-time vs batch conflation.
software-architecture
Design scalable software systems with proven architectural patterns (MVC, microservices, event-driven), SOLID principles, system design trade-offs, and architectural decision records (ADRs).