building-copilot-features

building-copilot-features is a skill for Claude Code, Codex from jovandyaz/knowtis-app. It costs 135 tokens per session (904 once invoked), scanned A, original, MIT.

A set of instructions for changing Knowtis's copilot and AI features, including its agent, model gateway, approval flow, safety checks, retrieval, memory, and evaluations. It defines rules for how messages, retrieved content, and changes must be handled.

In plain words
What is it for?
Adding agent tool groups, changing AI routing or fallback providers, implementing approval-required changes, maintaining prompt-injection protection, working with memory or retrieval, and running AI evaluations.
Why use it?
It helps developers avoid unsafe AI behavior, such as trusting user-supplied conversation history, applying changes without approval, or treating retrieved text as instructions.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one. Also seen: installed under .agents/ (shared by several agents).

Install

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.

agentmods
npx agentmods add skills/jovandyaz/knowtis-app/building-copilot-features
Any agent
npx skills add jovandyaz/knowtis-app --skill building-copilot-features
Clone the repo
git clone --depth 1 https://github.com/jovandyaz/knowtis-app

Made for: Claude Code, Codex.

Wrote 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.

agentmods badge for building-copilot-features

README.md
[![agentmods](https://agentmods.dev/badge/skills/jovandyaz/knowtis-app/building-copilot-features.svg)](https://agentmods.dev/skills/jovandyaz/knowtis-app/building-copilot-features)
Your own site
<a href="https://agentmods.dev/skills/jovandyaz/knowtis-app/building-copilot-features"><img src="https://agentmods.dev/badge/skills/jovandyaz/knowtis-app/building-copilot-features.svg" alt="Measured on agentmods" height="20"></a>
Per session 135 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 904 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00135 $0.00904
Opus 5 $0.00068 $0.00452
Sonnet 5 $0.00027 $0.00181
Haiku 4.5 $0.00014 $0.00090

Measured 6d ago against content hash 57db3efd01fe, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

building-copilot-features 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 6d 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.

.agents/skills/building-copilot-features/SKILL.md · 28 lines

How it starts

The opening of the file, as written. The whole thing — 28 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Building copilot features

The copilot is a server-authoritative conversational agent (apps/api/src/modules/agent/) on top of the framework-free @knowtis/ai-gateway. Safety invariants below are load-bearing — a change that violates one is a bug even if tests pass. Details: references/safety-invariants.md, references/module-map.md, references/eval-workflow.md.

Invariants you must never break

  1. Server-authoritative turns. The client sends one message + conversationId; the server rebuilds history from Postgres. Never trust client-supplied history.
  2. HITL on every mutation. Mutation tools emit a proposal parked in Redis; nothing is applied until agent:approve runs ApproveMutationHandler. No code path — including MCP — may auto-apply.
  3. Everything retrieved is DATA. Notes, memories, and web results are injected as data, never as instructions. detectPromptInjection() blocks at score ≥ 0.6; inputs > 50k chars are rejected (ReDoS defense). Every long-term-memory candidate fact passes the guard too.
  4. Fallback chain semantics. Primary → AI_FALLBACK_CHAIN, skipping keyless/cooldown providers. A stream never switches models mid-stream; aborts never advance the chain; the reported model is the one that actually served. BYOK turns skip the chain entirely and redact provider errors.
  5. BYOK encryption key is permanent. Keys are AES-256-GCM under one unversioned BYOK_ENCRYPTION_KEY; replacing it makes stored credentials appear missing. There is no lossless rotation path today. If compromise forces rotation, disable agent_byok, rotate the key, and require users to delete and re-enter provider keys before re-enabling it. Preserving credentials would require a separately designed dual-key, versioned re-encryption migration. BYOK bypasses the daily budget but still enforces RPM.
  6. Feature-flag gating pattern. Every AI feature is DB-flag gated (default off) and keyed to an env var: no VOYAGE_API_KEY → retrieval reconcile cron no-ops; no key → keep the flag off. "Feature not working" usually means a flag or key, not a bug.

Read the full file on GitHub · 28 lines

Files

What ships with it

4 files 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.

Changes

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.

  1. 6d ago First seen · 28 lines · 135 tokens per session scan A 57db3efd01fe

Subscribe to this mod's changes

building-copilot-features is a skill published in the GitHub repository jovandyaz/knowtis-app (3 stars, last pushed today), licensed MIT. It adds 135 tokens to every session and 904 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.

Related

Other skills, from other repositories