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.
git clone --depth 1 https://github.com/manishiitg/coding-agent-loopWrote 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/commands/manishiitg/coding-agent-loop/characters)<a href="https://agentmods.dev/commands/manishiitg/coding-agent-loop/characters"><img src="https://agentmods.dev/badge/commands/manishiitg/coding-agent-loop/characters/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/commands/manishiitg/coding-agent-loop/characters"><img src="https://agentmods.dev/badge/commands/manishiitg/coding-agent-loop/characters.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 | $0.00000 | $0.00606 |
| Opus 5.5 | $0.00000 | $0.00242 |
| Sonnet 5.5 | $0.00000 | $0.00121 |
| Haiku 4.5 | $0.00000 | $0.00061 |
Grade A, and why
characters 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 — 55 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Handle character work for this production — designing a new one, reviewing what exists, or revising a character already in use. Figure out which of these I mean from what follows, and if it's genuinely ambiguous, ask.
{{context}}
If no characters exist yet, or I'm asking to add one
Walk me through designing it, following video-cinematography's character
rules rather than skipping to a generation call:
- Establish what makes this character recognizable — 3-6 specific, disambiguating visual attributes (not "a woman in her 30s", but the exact details that would let someone tell her apart from any other character in this production). This becomes the one phrase reused verbatim in every later shot, so get it concrete now rather than refining it after shots already exist.
- Tell me which model and provider you're proposing to commit this
character's whole arc to, and why, per
video-model-selection— this is a real decision, not a default, and switching later is expensive. - Write the spec and generate the reference image, then call
show_characterand wait for me to approve it before generating any shot that uses it. An unapproved face propagates through the entire piece — this is the cheapest point in the whole production to catch it wrong.
If characters already exist and I'm asking to see them
Call show_character again for each one so they're current in the
Production panel, and tell me for each: which model and provider its arc is
committed to, and — this is usually the more useful part — everything about
how it's actually being used: which shots reference it, whether any of
those shots are already generated, and whether its spec or reference image
has drifted from what those shots were conditioned on.
If I'm asking to change an existing character
This is the expensive path, so be explicit about the cost before doing it.
Per video-cinematography, switching a character's model or provider
mid-arc is a last resort requiring my explicit sign-off — tell me how many
already-generated shots used the old version and would need to be
regenerated to match, not just that a change is possible. If I only want to
adjust the written spec (not the reference image or the model), that's
cheaper — say so, and say exactly which existing shots would then be out of
sync with the updated spec. Regenerate the reference image only after I've
confirmed I want the change, then show_character the result before it's
used in anything new.
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 · 55 lines · 0 tokens per session scan A 0554c919bd0a
characters is a command published in the GitHub repository manishiitg/coding-agent-loop (13 stars, last pushed 19d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 606 tokens. 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 2026-08-30.
Other commands, from other repositories
go
You MUST start your FIRST response with this exact agent activation line.
dashboard-session-tell
Send a prompt to another session. Usage /dashboard:session-tell.
dashboard-session-fork
Fork a session into a new one. Usage /dashboard:session-fork.
dashboard-session-hide
Hide a session from the dashboard list. Usage /dashboard:session-hide.
orchestrate
You MUST start your FIRST response with this exact agent activation line.
protocol-orchestration
You MUST start your FIRST response with this exact agent activation line.