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 ourobx/agent-ebpf --skill imaginegit clone --depth 1 https://github.com/ourobx/agent-ebpfWrote 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/ourobx/agent-ebpf/imagine)<a href="https://agentmods.dev/skills/ourobx/agent-ebpf/imagine"><img src="https://agentmods.dev/badge/skills/ourobx/agent-ebpf/imagine/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/ourobx/agent-ebpf/imagine"><img src="https://agentmods.dev/badge/skills/ourobx/agent-ebpf/imagine.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 | $0.00111 | $0.02896 |
| Opus 5.5 | $0.00044 | $0.01158 |
| Sonnet 5.5 | $0.00022 | $0.00579 |
| Haiku 4.5 | $0.00011 | $0.00290 |
Grade A, and why
imagine 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 1mo 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 — 202 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Imagine
Grok Build uses the split Imagine computer stack (grok_computer variants).
There is no consolidated imagine_image or imagine_video tool — always
call the modality-specific name from the table below.
| Tool | Role |
|---|---|
imagine_text_to_image |
New image from a text prompt only (no source). |
imagine_image_to_image |
Edit / restyle one existing image (sandbox path). |
imagine_reference_to_image |
Combine 2+ reference images (sandbox paths). |
imagine_text_to_video |
New video from a text prompt only (no source). |
imagine_image_to_video |
Animate one still (sandbox path) into a video. |
imagine_reference_to_video |
Video from 1+ reference images (sandbox paths). |
render_file |
Show a sandbox image/video path to the user in chat. |
Path-based handles. These tools read/write the shared sandbox:
- Generation returns a sandbox
file_path(under artifacts). Open it withread_file/ shell; show the user withrender_file. - Edit / animate tools take that path (or paths) as input — match the live
schema (
image,images, etc.). Never invent paths. - Never call
imagine_image/imagine_video. Asset-id helpers (imagine_create_asset/imagine_view_media/render_imagine_media) are not on this stack.
Apply this whenever you're considering or about to call any of these tools.
Handle flow (mandatory mental model)
generate → file_path → render_file (show user) / read_file or scripts (QC)
edit / animate → pass prior file_path(s) into image_to_* / reference_to_*
- Use the path the tool actually returned — do not invent filesystem paths.
- Show the user media with
render_file, not ad-hoc markdown image links.
Build accurate visuals with code, not the image tools
- Image models are unreliable at exact text, numbers, and structure. They can handle short text or a simple layout, but they often garble words, invent numbers, draw chart bars that match no data, or point diagram arrows nowhere, and the more that has to be exact, the worse they do. A detailed prompt doesn't make it dependable, and another
imagine_image_to_imageedit usually won't fix it. So when a result needs specific text, data, or structure to be correct (charts from real numbers, labeled or technical diagrams, math explainers, tables, screens with real copy), construct the asset with code, where you control the exact content. Prefer HTML and CSS, which give much better layout, typography, and polish than Python plotting. When only the look matters (photos, illustrations, characters, scenes, decorative art), the image tools are the right choice. Which one fits depends on what the output needs to get right, not on how the request is worded.
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.
- 1mo ago First seen · 202 lines · 111 tokens per session scan A c4ce1e170e2d
imagine is a skill published in the GitHub repository ourobx/agent-ebpf (1 stars, last pushed 1mo ago), licensed MIT. It adds 111 tokens to every session and 2,896 once invoked, about $0.0004 per session on Opus 5.5. 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 2026-09-04.
Other skills, from other repositories
fmc-firewall-ops
Cisco Secure Firewall FMC — access policy search, rule inspection, FTD device targeting, multi-FMC profile management. Use when searching firewall rules by IP or FQDN, checking if host A can reach host B through the firewall, auditing FMC access policies, or reviewing SGT-based segmentation rules.
fortianalyzer-ops
FortiAnalyzer log operations — policy-filtered traffic log query within a bounded time window, offset pagination, per-policy activity checks, and logging-device inventory. Use when asking whether traffic actually matched a firewall rule, investigating what hit a policy, or determining whether a rule is genuinely…
fortigate-ops
FortiGate device operations — system status, interfaces, routing, IPsec VPN tunnel state with phase 1 and phase 2 reported separately, HA member identification, per-VDOM scoping, and manager-vs-device drift detection. Use when asking what a FortiGate is ACTUALLY doing right now, whether a tunnel is up, or whether the…
fortimanager-ops
FortiManager policy operations — ADOM inventory, policy package review, recursive object resolution, revision history, install preview, and gated package install. Use when auditing FortiGate firewall policy at the MANAGER level (intent), reviewing ADOM policy packages, or planning a package install with rollback…
fwrule-analyzer
Multi-vendor firewall rule analysis — overlap detection, shadowing, conflict identification, duplication checking across PAN-OS, ASA, FTD, IOS/IOS-XE, IOS-XR, Check Point, SRX, Junos, Nokia SR OS, and Fortinet FortiOS/FortiGate. Use when validating firewall rule changes, auditing rulesets for conflicts, or normalizing…
azure-network-ops
Azure cloud networking -- VNets, NSGs, ExpressRoute, VPN Gateways, Azure Firewalls, Load Balancers, Application Gateways, Route Tables, Network Watcher, Private Endpoints, DNS zones. Use when auditing Azure VNets, troubleshooting hybrid connectivity (ExpressRoute/VPN), inspecting what an NSG currently allows…