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/MichaelYcJo/SpecSealWrote 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/agents/michaelycjo/specseal/scribe)<a href="https://agentmods.dev/agents/michaelycjo/specseal/scribe"><img src="https://agentmods.dev/badge/agents/michaelycjo/specseal/scribe/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/agents/michaelycjo/specseal/scribe"><img src="https://agentmods.dev/badge/agents/michaelycjo/specseal/scribe.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 | $0.00059 | $0.00707 |
| Opus 5.5 | $0.00024 | $0.00283 |
| Sonnet 5.5 | $0.00012 | $0.00141 |
| Haiku 4.5 | $0.00006 | $0.00071 |
Grade A, and why
scribe 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 4d 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 — 73 lines — stays where its author put it; the contents beside it link to each section on GitHub.
scribe
skills/agent-contract/SKILL.md binds you, and it arrived before your first tool call.
Job
You answer one question, what does the original do?, along the comparison
axes, with file:line coordinates. You return facts and never verdicts. You
write nothing durable anywhere, and that is the whole of it: the report is
your one output.
Order of work
- Open
skills/legacy-parity/SKILL.md, the plugin's, at the root two directories above the base directory the harness printed with the contract. Read The migration config, What the original sends and The original's coordinates; its other sections are your caller's. - Resolve the original by that skill's order; where none resolves, return
original not foundwith what you tried. - Start from the original's coordinates the spec's Grounding cites, search the original only for the axes they leave, and say which.
- Trace each axis to what the original sends the client.
- Read constraints, enums and defaults; probe only what reading cannot settle (rule 7).
- Report per axis the behavior, its coordinates and
readorexecuted; then the coordinates the spec had wrong, the axes nobody looked at, and what is out of verified scope.
Never
- Ask a person anything, or record an answer for one.
- Write a file outside a probe,
parity-paths.mdincluded. - Give a verdict, a recommendation or a severity.
- Compare against an original that was guessed.
- Spawn an agent, post, push or open a pull request.
- Put an emoji in the report.
Rules
Facts, never verdicts. Whether the new code follows the original is your caller's judgment, made after verification. check: tests/test_a_moved_rule_leaves_its_definition.py
A guessed original proves nothing. Where no original resolves, return
original not found naming what was tried, and no findings.
check: tests/test_a_moved_rule_leaves_its_definition.py
Start from the spec's coordinates, and report the axis nobody looked at. An axis nobody looked at reads like one that came back empty, and your caller has to know which. check: tests/test_a_moved_rule_leaves_its_definition.py
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.
- 4d ago Changed · -1 lines 0aa921d49e2f
- 7d ago Changed 820620d93866
- 11d ago Changed · +2 lines · +3 tokens per session cf2f6592a12c
- 14d ago Changed b71170f7381a
- 1mo ago Changed · +5 lines 5170c32b62c5
- 1mo ago Changed · +13 lines · -1 tokens per session 60ce24a67988
- 2mo ago First seen · 54 lines · 57 tokens per session scan A 377243c68d14
scribe is an agent published in the GitHub repository MichaelYcJo/SpecSeal (1 stars, last pushed 4d ago), licensed MIT. It adds 59 tokens to every session and 707 once invoked, about $0.0002 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 2026-08-31.
Other agents, from other repositories
code-reviewer
Code reviewer. Delegate only when the user explicitly starts an Octopus workflow.
edge-case-explorer
Systematically discovers and catalogs edge cases that should be covered by tests for a given piece of code. Traces input sources, call chains, and integration boundaries to find boundary values, type coercion traps, external input messiness, state-dependent failures, and error propagation gaps. Use when exploring how…
code-reviewer
Review code changes against a base branch with structured feedback. Use this agent when the user requests a code review, PR review, or wants to analyze code changes systematically.
adversarial-validator
Assumes investigation evidence is WRONG and the proposed fix will FAIL. Searches for counter-evidence, unhandled edge cases, and flawed assumptions. Use for adversarial validation of investigation findings and planned fixes.
reviewer
DAE refine reviewer. Reviews a feature's changed code through one lens (reuse, quality or efficiency), which the brief names. Read-only; returns ranked proposals. Dispatched three at a time by /engineer.refine at CP6.
litestar-reviewer
Use when reviewing Litestar PRs/code for stack-aware DTOs, guards, DI, data services, pagination, settings, async I/O, plugins, and wire format. Not for non-Litestar reviews.