docs: reflect v2 pivot — agent hosting, credential firewall

[indykish]

Summary

• index.mdx: Updated to v2 positioning. Added "Why we pivoted" rationale (from README). Replaced spec-to-PR feature cards with agent hosting features: always-on agents, credentials hidden, webhooks wired, observability, spend ceiling, kill switch. Removed in-progress pivot notice from Tip.
• how-it-works.mdx: Full rewrite. Spec-to-PR lifecycle replaced with agent hosting model — credential firewall, always-on execution, webhook routing, observability. Includes mermaid diagrams for the hosting overview, credential firewall sequence, and runtime architecture.

Test plan

• Verify docs site renders correctly (Mintlify MDX components, mermaid diagrams)
• Check all internal links still resolve
• Confirm pivot rationale reads consistently with docs(readme): reflect product pivot, fix logo usezombie#168

🤖 Generated with Claude Code

[greptile-apps]

Greptile Summary

This PR rewrites index.mdx and how-it-works.mdx to reflect UseZombie's v2 pivot from a spec-to-PR code-generation tool to an agent hosting platform with a credential firewall. The changes are well-structured and the previous P1 review findings have all been resolved in this commit.

• index.mdx: Adds a "Why we pivoted" rationale section and replaces the old feature cards with six new ones covering always-on agents, credential hiding, webhooks, observability, spend ceiling, and kill switch. Navigation cards are updated to point to the new agent hosting docs.
• how-it-works.mdx: Full rewrite introducing the credential firewall model with three Mermaid diagrams (hosting overview, firewall sequence, runtime architecture) and a five-step onboarding walkthrough. Previous P1 issues — wrong CLI command (zombiectl secrets add → zombiectl skill-secret put), inconsistent "Mission Control" terminology, and missing backtick formatting on zombiectl/zombied — are all fixed in this commit.
• README.md: Minor tagline update and copyright year change; no issues.
• One small style nit: index.mdx line 23 uses "in minutes," a time estimate disallowed by the AGENTS.md style guide.

Confidence Score: 5/5

Safe to merge — all prior P1 concerns resolved; one trivial P2 style nit remains.

All previous P1 issues (wrong CLI command, missing Mission Control branding, missing code formatting on zombiectl/zombied) are addressed in this commit. The only remaining finding is a P2 wording issue ('in minutes' time estimate) that does not affect documentation correctness or user paths.

No files require special attention.

Important Files Changed

Filename	Overview
index.mdx	Rewrites introduction to v2 agent-hosting pitch — feature cards and nav updated; one minor time-estimate wording nit on line 23.
how-it-works.mdx	Full rewrite to agent-hosting model with three Mermaid diagrams; all prior P1 CLI, terminology, and formatting issues resolved.
README.md	Minor update — revised tagline and copyright year; no issues.

Sequence Diagram

sequenceDiagram
    participant Dev as Developer
    participant CLI as `zombiectl`
    participant API as `zombied` API
    participant Q as Redis Streams
    participant W as `zombied` worker
    participant S as Sandbox process
    participant F as Credential firewall
    participant E as External API

    Dev->>CLI: deploy agent
    Dev->>CLI: zombiectl skill-secret put
    CLI->>API: POST /v1/agents (agent config)
    API->>Q: enqueue agent_start
    Q->>W: claim work
    W->>S: spawn sandboxed process
    S->>F: outbound request (no credentials)
    F->>F: lookup(workspace_id, target_host)
    F->>E: forwarded request + injected Bearer token
    E-->>F: 200 OK
    F-->>S: 200 OK (credential never in sandbox)
    F->>F: audit_log.append(action, ts, cost_tokens)
    W->>CLI: SSE: status, logs, cost

Loading

Prompt To Fix All With AI

This is a comment left during a code review.
Path: index.mdx
Line: 23

Comment:
**Time estimate in user-facing copy**

The phrase "in minutes" is a time estimate, which conflicts with the AGENTS.md rule: *Do not use time estimates or effort ratings in user-facing docs*. Actual setup time varies across agent complexity, credential count, and webhook configuration, so this phrasing could also feel like an unfulfilled promise.

```suggestion
  Connect your agent and have it running quickly.
```

**Context Used:** AGENTS.md ([source](https://app.greptile.com/review/custom-context?memory=e994989b-b25e-4468-be5d-6bde3ddb0ce6))

How can I resolve this? If you propose a fix, please make it concise.

_{Reviews (2): Last reviewed commit: "fix: address greptile P1/P2 — correct CL..." | Re-trigger Greptile}