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/perniemann/pncore/pn-context-engineeringnpx skills add perniemann/pnCore --skill pn-context-engineeringgit clone --depth 1 https://github.com/perniemann/pnCoreWrote 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/perniemann/pncore/pn-context-engineering)<a href="https://agentmods.dev/skills/perniemann/pncore/pn-context-engineering"><img src="https://agentmods.dev/badge/skills/perniemann/pncore/pn-context-engineering.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 | $0.00045 | $0.01144 |
| Opus 5 | $0.00023 | $0.00572 |
| Sonnet 5 | $0.00009 | $0.00229 |
| Haiku 4.5 | $0.00005 | $0.00114 |
Grade A, and why
pn-context-engineering 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 2d 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 — 88 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Context engineering
When to use
- Starting a coding session or onboarding an agent to a repo
- Switching tasks or domains mid-session
- Output drifts (wrong APIs, ignored conventions, invented paths)
- Token budget pressure: need to slim what is loaded without losing guardrails
Context hierarchy (most → least persistent)
- Rules and house docs —
.cursor/rules/,AGENTS.md,CLAUDE.md, project README conventions (always-on constraints). - Specs and architecture — discovery spec, PRD, ADRs, plans (
docs/plans/,docs/discovery/) for the current initiative. - Task-scoped sources — only files that implement or verify the active change.
- Ephemeral signals — test output, build logs, linter diagnostics, stack traces (refresh each iteration; do not treat stale logs as ground truth).
- Conversation — accumulates; prefer writing decisions back to spec/plan rather than relying on long threads.
Process
- State what matters now: one sentence on goal and non-goals for this step.
- Load tier 1 if missing: ensure rules/agents reflect repo truth (commands, paths, forbidden actions). Load a skill once per session; do not restack the same rule dump every turn (breaks prompt cache and, on Fable 5.1, thinking-block binding).
- Pull tier 2 only for the active feature — not the whole docs tree.
- Expand tier 3 surgically: open callers, contracts, and tests tied to the edit; avoid "read the whole package."
- After each failure: replace guesswork with tier 4 evidence; if evidence contradicts the plan, update the plan/spec (orchestration philosophy: map is not the territory).
Rationalizations
| Excuse | Reality |
|---|---|
| "I'll paste the whole repo into context." | Noise hides the relevant files; use search and targeted reads. |
| "Rules are enough; I don't need the spec." | Rules don't encode acceptance criteria or this week's scope. |
| "The user already explained it in chat." | Chat compacts and drifts; canonical spec/plan wins. |
| "One more huge file won't matter." | It crowds out signal; load slices or symbols, not megabytes. |
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 2d ago First seen · 88 lines · 45 tokens per session scan A bb5ed79e51f2
pn-context-engineering is a skill published in the GitHub repository perniemann/pnCore (0 stars, last pushed 2d ago), licensed MIT. It adds 45 tokens to every session and 1,144 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 2026-09-03.
Other skills, from other repositories
wisdom-forget
Remove a must-keep CocoWisdom entry with rationale. Usage: $wisdom forget --id --rationale " ".
cocoplus-console
Launch CocoConsole, the CocoPlus 2.0 read-only local browser control plane. Use $cocoplus console, $cocoplus console status, or $cocoplus console stop.
honey-ccr
Compress-cache-retrieve huge array tool output.
screenpipe-api
Query the user's local and synced-device Screenpipe data via the REST API at localhost:3030. Use for screen activity, meetings, apps, productivity, other-device or cross-device history, media export, retranscription, or connected services.
meter-waste
Analyze CocoMeter context waste and surface repeatable reduction opportunities.
pulse-off
Disable CocoPulse ambient progress observation.