Getting it into your agent
This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.
/plugin marketplace add rjmurillo/ai-agents/plugin install project-toolkitWrote 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/rjmurillo/ai-agents/orphan-ref-validator)<a href="https://agentmods.dev/skills/rjmurillo/ai-agents/orphan-ref-validator"><img src="https://agentmods.dev/badge/skills/rjmurillo/ai-agents/orphan-ref-validator/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/rjmurillo/ai-agents/orphan-ref-validator"><img src="https://agentmods.dev/badge/skills/rjmurillo/ai-agents/orphan-ref-validator.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 | $0.05829 |
| Opus 5.5 | $0.00036 | $0.02332 |
| Sonnet 5.5 | $0.00018 | $0.01166 |
| Haiku 4.5 | $0.00009 | $0.00583 |
Grade E, and why
orphan-ref-validator scanned grade E with 2 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.
Hidden instructionshighPrompt injection
Directives inside HTML comments, invisible characters or bidirectional overrides are read by the model and not by the person reviewing the file.
| `<!-- orphan-ref-ignore-file -->` | Whole file | Anywhere in the **first 50 lines** of the file | Skip the file entirely; emit no findings and report suppressed references under `directive_suppressed`. | Reaches for credential fileshighPrivilege escalation
SSH keys, cloud credentials, git-credentials, .npmrc, /etc/shadow: reading these is how a config file becomes a credential leak.
Target paths are resolved with `pathlib.Path.resolve()` and must lie under the repository root. Paths outside the repo, broken symlinks, symlink loops, outside-repo symlinks, walk errors, stat errors, unreadable files, a How it starts
The opening of the file, as written. The whole thing — 392 lines — stays where its author put it; the contents beside it link to each section on GitHub.
orphan-ref-validator
Purpose
Scans structured artifacts (specs, ADRs, eval fixtures, plugin manifests, skill descriptions) for references to entities that do not exist in the working tree:
- Skill names that no longer have a
.claude/skills/<name>/directory. Emitted asFinding(kind="skill_name", severity="critical"). - Script paths under
build/scripts/,scripts/validation/,scripts/, ortests/that are not present on disk. Emitted asFinding(kind="script_path", severity="critical"). - Rule and instruction paths under
.claude/rules/,.github/instructions/, orsrc/copilot-cli/instructions/that are not present on disk. Emitted asFinding(kind="rule_path" | "instruction_path", severity="critical"). - Retired-surface paths: a citation of the slash-command directory that ADR-064 retired, in live guidance documents only (skills, rules, instruction mirrors; not source code). Emitted as
Finding(kind="retired_path", severity="critical"). A line that saysMigrated fromor names ADR-064 records the retirement and is not flagged. Specs, ADRs, and tests are dated records and are not checked.
Emits findings per the ADR-056 envelope and a final verdict line. Exit code follows ADR-035: VERDICT: PASS or VERDICT: WARN exits 0; VERDICT: CRITICAL_FAIL exits 1; usage/configuration or incomplete scan exits 2; external failures exit 3; permission failures exit 4.
The skill ships with vendored installs. Missing targets are errors by default. Use --allow-missing-targets only when the caller intentionally runs in a partial vendored tree, and use --allow-empty-scan only when zero scanned files is the declared scope.
Triggers
| Trigger | Effect |
|---|---|
scan for orphan refs |
Run with default targets |
validate orphan references |
Run on a specific path |
check skill catalog drift |
Run with default targets |
build mandatory exit gate |
Invoked by the build lifecycle command |
Path conventions
What ships with it
7 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.
- 4d ago Changed · +2 lines 382435ebd587
- 9d ago Changed 3fc02d9b5e48
- 18d ago Changed · +8 lines 4e22d313c073
- 20d ago Changed ce7259ac343c
- 22d ago Changed · +6 lines 1f0d39dcc44b
- 24d ago Changed 0ffb68f80a12
- 1mo ago Changed · +1 lines bb97f98ab993
- 1mo ago First seen · 375 lines · 90 tokens per session scan E 69f9c7a8e800
orphan-ref-validator is a skill published in the GitHub repository rjmurillo/ai-agents (48 stars, last pushed today), licensed MIT. It adds 90 tokens to every session and 5,829 once invoked, about $0.0004 per session on Opus 5.5. A static security scan graded it E with 2 findings (hidden instructions, reaches for credential files). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
review-remote
Review open PR/MR and publish findings.
extension-audit-methodology
Systematic 5-dimension audit framework for VS Code extensions — debug hygiene, dead code, performance, menu validation, dependency cleanup.
github-automation
GitHub workflow automation, PR management, issue tracking, and code review coordination. Integrates with GitHub Actions and repository management. Use when: PR creation, code review, issue management, release automation, workflow setup. Skip when: local-only changes, non-GitHub repositories.
git-advanced-workflows
Master advanced Git workflows including rebasing, cherry-picking, bisect, worktrees, and reflog to maintain clean history and recover from any situation. Use when managing complex Git histories, collaborating on feature branches, or troubleshooting repository issues.
mcore-create-issue
Investigate a failing GitHub Actions run or job and create a GitHub issue for the failure.
review-this-branch
Run the nohuman review gate (fresh-session adversarial reviewer + tamper guard) over the current branch or a GitHub pull request, with no server, no database, and no onboarding, and relay the pass/fail checklist.