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 bks-lab/open-bridge --skill bridge-auditgit clone --depth 1 https://github.com/bks-lab/open-bridgeWrote 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/bks-lab/open-bridge/bridge-audit)<a href="https://agentmods.dev/skills/bks-lab/open-bridge/bridge-audit"><img src="https://agentmods.dev/badge/skills/bks-lab/open-bridge/bridge-audit.svg" alt="Measured on agentmods" 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 | $0.00167 | $0.02595 |
| Opus 5 | $0.00084 | $0.01298 |
| Sonnet 5 | $0.00033 | $0.00519 |
| Haiku 4.5 | $0.00017 | $0.00260 |
Grade A, and why
bridge-audit 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 7d 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 — 174 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Bridge Audit — Drift Detection
bridge-audit is the systematic version of "let me re-read the README
and see what doesn't match anymore". It runs 13 categorical checks
against the current repo state and returns a stratified report.
Check 9 only runs with --cross-repo (it fetches sister-repo trunks);
Check 10 (agent-identity health), Check 11 (gate-shaped memory),
Check 12 (config-driven CORE skills), and Check 13 (user-level skill
shadowing) always run.
Read the referenced file ONLY when triggered.
Arguments
| Argument | Effect | Default |
|---|---|---|
(none) |
Full audit, all checks | — |
--check <name> |
Run only one check (license / skill-tree / protocol-count / renames / xrefs / scopes / routing-sot / typos / agent-identity / memory-gate / config-driven / skill-shadowing) | all |
--cross-repo |
Also clone configured upstreams and compare README/AGENTS for divergence | false |
--strict-oss |
When run on an OSS variant: flag hardcoded internal vocabulary (delegates to bridge-leak-check) |
false |
--fix |
Where unambiguous, apply the suggested fix | false (advisory only) |
--no-history |
Skip writing the post-run JSON to work/_learning/audit-history/ |
history written by default if learning.enabled |
--no-recurring-scan |
Skip the trend-detection step that auto-generates proposals for recurring findings | scan runs by default |
The 13 Checks
| # | Check | What it compares | Severity hint |
|---|---|---|---|
| 1 | License consistency | README badge ↔ LICENSE file ↔ README footer ↔ CLAUDE.md license claim ↔ MEMORY mentions | P0 if any disagree |
| 2 | Skill-tree truth | README skill table/tree ↔ AGENTS.md skill table ↔ skills/*/SKILL.md actual |
P1 |
| 3 | Standing-order count | "N standing orders" claims ↔ protocols/standing-orders/*.md count (excluding _template.md, README.md) |
P1 |
| 4 | Renamed-everywhere | data/renames.yaml old-name greps in tracked content (excluding listed exceptions) |
P1 |
| 5 | Cross-reference validity | Markdown [text](path) and `path` references resolve to existing files |
P2 |
| 6 | Scope coverage | Group A: skills/*/SKILL.md have explicit scope: under metadata: (metadata.scope); .claude/agents/*.md have explicit top-level scope: (P2); rules/*.md scope is REQUIRED + hard-gated by validate-bridge.py (P1). Group B: every identity/{mandants,accounts,personas}/*.yaml + workflow/{contexts,projects}/*.yaml + infra/{remotes,channels}/*.yaml has the required top-level scope: (P1 — schema-enforced, routing-critical) |
P1/P2 |
| 7 | Routing-SoT conflicts | Multiple files claim authoritative status for the same routing domain (heuristic: same path appears in multiple "source of truth" tables with different rules) | P2 |
| 8 | Typo lint | Bridge-corpus typo patterns from data/typo-patterns.yaml |
P3 |
| 9 | Skill-tree sync coverage (--cross-repo only) |
Local skills/*/ scope ↔ each upstream's skills/ directory listing — forward-drops + reverse-leaks |
P1 (P0 if outstanding >2 sync windows) |
| 10 | Agent-identity health | identity/agent/SOUL.md present + ≤80 lines / ≤4 KB · IDENTITY.md present · both carry valid frontmatter (schema_version/type/scope/last_updated) |
P1 if SOUL.md missing, else P2 |
| 11 | Gate-shaped memory without a rules/ home |
Memory files whose body uses gate language (imperative + always/never/immer/nie, or "when X → do Y" routing) but have NO corresponding rule in rules/ (core/bks/user) — a behavioral rule trapped in the private store |
P2 |
| 12 | Config-driven CORE skills | scope: core skill files (SKILL.md + references/) that hardcode instance specifics — org/project IDs, tracker queries, persona names, pipeline IDs, absolute instance paths — instead of reading them from bridge-config.yaml / workflow/ / infra/ / identity/ (CLAUDE.md § Generic CORE Skills) |
P2 |
| 13 | User-level skill shadowing | ~/.claude/skills/ (resolved) ↔ this repo's skills/ — the user level overrides the project level, so a pointer at a Bridge repo silently serves one instance's skills to every other instance (AGENTS.md § Skills). Both directions: this instance shadowing others, and others shadowing this one (with diff -rq drift per colliding name). Also reports scripts that resolve the path as a filesystem location |
P0 if own skills are shadowed or instance-bound skills (scope org or user) leak machine-wide, else P1 |
What ships with it
4 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.
- 7d ago First seen · 174 lines · 167 tokens per session scan A 65e95ca28430
bridge-audit is a skill published in the GitHub repository bks-lab/open-bridge (8 stars, last pushed today), licensed MIT. It adds 167 tokens to every session and 2,595 once invoked, about $0.0008 per session on Opus 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 skills, from other repositories
langsmith-tracing
LangSmith tracing and debugging setup for LLM applications. Configure observability, capture traces, and enable debugging for LangChain/LangGraph agents.
behavior-contract
Bug condition/postcondition formalization as testable Behavior Contracts. Defines invariants that must be preserved across fixes.
quality-hooks
Language-specific auto-lint/format/typecheck pipeline. Supports Python (ruff+pyright), TypeScript (prettier+eslint+tsc), Go (gofmt+golangci-lint). Auto-fix and convergence loops.
systematic-debugging
Structured debugging methodology using hypothesis-driven investigation, log analysis, and bisection to isolate and resolve defects.
debugging-patterns
Root cause analysis frameworks including log-first investigation, git bisect correlation, and pattern-based diagnosis with confidence scoring.
hotfix-triage
Urgent issue classification, root cause analysis, and fast-path routing for production hotfixes.