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 forefy/.context --skill evidence-videogit clone --depth 1 https://github.com/forefy/.contextWrote 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/forefy/.context/evidence-video)<a href="https://agentmods.dev/skills/forefy/.context/evidence-video"><img src="https://agentmods.dev/badge/skills/forefy/.context/evidence-video/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/forefy/.context/evidence-video"><img src="https://agentmods.dev/badge/skills/forefy/.context/evidence-video.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 | $0.00090 | $0.02929 |
| Opus 5.5 | $0.00036 | $0.01172 |
| Sonnet 5.5 | $0.00018 | $0.00586 |
| Haiku 4.5 | $0.00009 | $0.00293 |
Grade A, and why
evidence-video 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.
How it starts
The opening of the file, as written. The whole thing — 91 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Evidence video
Build a narrated-by-structure video from real captures only. The generator refuses to emit a scene without a real frame behind it, each step hashes the one before it, and the closer explains what the flow means. This is for proving a reproduction, walking a reviewer through an auth or access flow, or showing a user story end to end.
Two hard rules the tool enforces, and you must not work around:
- No scene without a real capture. Every step is a screenshot the tool took, hashed into a manifest. Never hand-draw a UI, never fabricate a frame, never claim a step you did not capture.
- Never fabricate on-screen content. Every URL, path, status code, number, label, and request or response line shown on the video must come from a real capture or a real source file. If you do not have the real value, leave it out. Do not invent a plausible URL, a round-number timing, a status, an endpoint, or a label to fill a gap. One made-up detail turns the whole thing from evidence into a mockup, and the hash chain cannot catch a fabricated caption, only a swapped frame, so this discipline is yours. When in doubt, show less.
- Keep the hash chain off the video. It is the integrity mechanism, not viewer content: the chain lives in the manifest and
verify, never as an on-screen scene or a hash printed on a frame. The video's job is the story and its meaning. - Always end with insight, impact and risk. It is the point of the video: who is affected, what is at stake, what to verify, and the blast radius. A video that stops at the walkthrough is unfinished.
Dependencies
python3(standard library only) forassets/evidence.py.- HyperFrames for rendering:
npx hyperframes(Node 22+).checkvalidates,renderwrites the MP4. - A capture source. Any one of:
- Burp proxy history over the
reburpREST API (http://127.0.0.1:9090) for the real HTTP request and response behind a step. See theburp-interactionusage if present. - A browser you can drive and screenshot (Claude in Chrome, or the built-in browser pane) for the UI frames and live network list.
- Terminal screenshots for a CLI or a Foundry or
forge testreproduction, where each step is a command and its real output.
- Burp proxy history over the
What ships with it
2 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.
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 Changed · +19 lines 36039648f3e7
- 3d ago First seen · 72 lines · 90 tokens per session scan A aaf03f0e37d6
evidence-video is a skill published in the GitHub repository forefy/.context (153 stars, last pushed yesterday), licensed MIT. It adds 90 tokens to every session and 2,929 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 2026-10-19.
Other skills, from other repositories
argus
Argus — the all-seeing scanner suite. Six automated scanners for high-value web + LLM bug classes — CORS misconfiguration (origin reflection / null / credentialed read), CRLF & host-header injection, NoSQL injection (operator auth-bypass / $where blind), JWT attacks (alg:none / RS256→HS256 confusion / secret crack)…
beehave
Use when using the Beehave addon — pure-GDScript behavior trees with composites, decorators, leaves, a blackboard, and a visual runtime debugger.
asc-crash-triage
Triage TestFlight crashes, beta feedback, and performance diagnostics using asc. Use when the user asks about TF crashes, TestFlight crash reports, beta tester feedback, app hangs, disk writes, launch diagnostics, or wants a crash summary for a build or app.
om-auto-fix-issue
Fix or implement a tracker issue end to end from a single command — takes an issue id or a plain problem description (filed first via om-prepare-issue), classifies, then drives the bug autofix chain (om-verify-in-repo, om-root-cause, om-fix, om-open-pr, om-auto-review-pr, om-auto-qa-pr for UI fixes) or the feature…
deliver-edge-cases
Documents edge cases, error states, boundary conditions, race conditions, and recovery paths for a feature - the systematic catalog of what can go wrong and the failure modes to design for. Use during specification to map the failure surface and ensure comprehensive coverage, or during QA planning to identify boundary…
om-fix
Implements the minimal code change identified by the om-root-cause step, adds regression tests, and runs the configured validation gate. Claims the tracker issue at start (assignee + in-progress label + claim comment) so concurrent automation backs off. Does not commit, push, or open a PR — that is the om-open-pr…