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/raccioly/docguardWrote 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/raccioly/docguard/anchors)<a href="https://agentmods.dev/commands/raccioly/docguard/anchors"><img src="https://agentmods.dev/badge/commands/raccioly/docguard/anchors/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/raccioly/docguard/anchors"><img src="https://agentmods.dev/badge/commands/raccioly/docguard/anchors.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 | $0.00011 | $0.00624 |
| Opus 5.5 | $0.00004 | $0.00250 |
| Sonnet 5.5 | $0.00002 | $0.00125 |
| Haiku 4.5 | $0.00001 | $0.00062 |
Grade A, and why
anchors 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 2d 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 — 65 lines — stays where its author put it; the contents beside it link to each section on GitHub.
DocGuard Anchors
List every @implements / @req tag in source code, the requirement it
resolves to, and the file, symbol, statement or route it anchors to, with the
tier that placed it. Read-only: this command never writes the anchor lock.
Running DocGuard
Run docguard from PATH when it is installed. Otherwise run
npx --yes [email protected], the release these instructions ship with, in its
place. Every docguard … command below means one of the two.
User Input
$ARGUMENTS
Steps
- List the links, scoped to the spec being worked on when there is one:
docguard anchors --format json $ARGUMENTS
-
For each
problems[]entry:- ANC007: the tag cites a requirement nobody declares. Qualify it with the Spec ID that declares it, or fix the ID.
- ANC006 (
severity: info): the tag's scope is ambiguous or did not resolve. Add{scope=symbol}or{scope=file}after the requirement, or move the tag directly above the code it describes. - ANC010: links were placed below their exact tier. Report it; do not edit code to silence it.
-
For each link whose code changed (
state: suspect,side: code), read itstests[]. A test markedneeds-reverificationshould be run, or its TestGuard claim probed. A claim names the requirement as"source": { "kind": "spec", "ref": "req:<spec-id>#<REQ>" }. Treatinconclusiveas unproven, never as passing. -
When implementing a requirement, put its tag directly above the function, class or statement that implements it, not in the file header, so the link names that code.
-
For code with no requirement, run
docguard anchors --orphans --format json. For each orphan, tag it with the requirement it implements. If none applies, tell the user which disposition fits (derived,deadorinfrastructure) and why. Recording one writes the anchor lock, which is the user's sign-off, not yours.
Flags
--spec <id>: only links to this Spec ID--file <path>: only links in this file--orphans: list untagged exported or top-level symbols instead of links--format json: machine-readable output
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.
- 2d ago Changed e92c172a2ec9
- 4d ago First seen · 65 lines · 11 tokens per session scan A 34fc498beb0a
anchors is a command published in the GitHub repository raccioly/docguard (32 stars, last pushed today), licensed MIT. It adds 11 tokens to every session and 624 once invoked, about $0.0000 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 2026-10-12.
Other commands, from other repositories
implement-evidence-asks
Read by /implement Phase 1 (D-CMD-REF-HAND). "The grammar below", "the exception grammar above" and "the record above" name sections of this file. Any other "above" means the Phase 1 text before the step that sent you here.
afe
Evaluate feature - code review or comparison (shortcut for feature-eval).
aitask-pr-import
Create an aitask from a pull request by analyzing PR data and generating a structured task with implementation plan.
aitask-trail
Create, refresh, or show an implementation trail — a durable, wave-structured, evidence-backed task-sequencing artifact stored via ait artifact.
aitask-review
Review code using configurable review guides, then create tasks from findings.
aitask-wrap
Wrap uncommitted changes into an aitask with retroactive documentation and traceability.