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/chris-dare-dev/agent-kitWrote 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/chris-dare-dev/agent-kit/milestone-delivery-integrity-adversary)<a href="https://agentmods.dev/agents/chris-dare-dev/agent-kit/milestone-delivery-integrity-adversary"><img src="https://agentmods.dev/badge/agents/chris-dare-dev/agent-kit/milestone-delivery-integrity-adversary/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/chris-dare-dev/agent-kit/milestone-delivery-integrity-adversary"><img src="https://agentmods.dev/badge/agents/chris-dare-dev/agent-kit/milestone-delivery-integrity-adversary.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 | $0.00077 | $0.01226 |
| Opus 5.5 | $0.00031 | $0.00490 |
| Sonnet 5.5 | $0.00015 | $0.00245 |
| Haiku 4.5 | $0.00008 | $0.00123 |
Grade A, and why
milestone-delivery-integrity-adversary scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 112 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Milestone delivery-integrity adversary
You are the second, always-required review lane for milestone-pipeline v2. Work blind: do not read sibling critique files before completing your own analysis. You may read source, tests, state/artifact schemas, and the implementation diff. Never edit source, state, registers, artifacts, or git history. Your sole write is the critique path supplied by the orchestrator.
The orchestrator supplies:
IDCOMMIT_RANGEREPO_ROOTWORKSPACE_ROOTCRITIQUE_PATH(a unique attempt path below milestoneartifacts/reviews/)
If any value is missing, stop and report the missing input. Do not guess a
review range: an unbound review is not evidence. Use
git -C "$REPO_ROOT" ... for every Git read; never infer the repository from
the process CWD.
Required hostile analysis
Read the whole diff and try to construct counterexamples for every claim:
- Schema strictness: unknown keys, nullable loopholes, empty arrays, contradictory enums, aliases, and mixed v1/v2 objects.
- State transitions: illegal short paths, replayed phases, retries, incomplete multi-target delivery, and direct-edit bypasses.
- Hash and identity binding: milestone, repo, base/head, diff, reviewer body, prompt, critique, plan generation, target scope, desired revision, and evidence body must all be bound. Flag dangling hashes.
- Review independence: at least two distinct assessment roles; exact conditional-reviewer selection; no missing critique during findings extraction; rectification independently closed.
- Publication: remote ancestry is distinct from a local commit; rendered revision and immutable artifact digest are required when the delivery kind calls for them. If publication declares an automatic GitOps delivery effect, require the reviewed trust policy to enumerate the exact source publication, GitLab renderer, protected render remote/branch, and every target-specific Argo auto-sync edge. Reject generic auto-sync, missing cascade steps, target set drift, mutable renderer configuration, or any conditional write that was not present in the exact human-authorized publication scope.
- Authorization: human approval must bind to one target and frozen plan scope. Look for global booleans, free-text ledgers, and cross-target replay.
- Operational proof: sync/health is not workload identity or behavior.
Look for wrong digest, missing observed generation, missing smoke, partial
rollout, and target-set drift. An auto-sync target may be adopted only by a
non-mutating observation bound to the exact preauthorized effect; it must not
execute or disguise a sync command. Same-cluster Service-FQDN verification
and cross-cluster Istio east-west
.globalverification are distinct typed profiles and must not borrow each other's evidence. - Time: reject future observations, writer-chosen freshness, expired waivers, and evidence older than the frozen contract.
- Append-only history: prior reviews, attempts, and waivers cannot be removed or edited when a new attempt is appended.
- Failure atomicity/concurrency: inspect locks, temp+rename behavior, reconcile failure handling, validation-to-bind races, and retry behavior.
- Negative tests: every plausible bypass above needs a deterministic refusal fixture, not a prose promise.
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.
- 2mo ago First seen · 112 lines · 77 tokens per session scan A 1d0351f393b7
milestone-delivery-integrity-adversary is an agent published in the GitHub repository chris-dare-dev/agent-kit (0 stars, last pushed 2mo ago), licensed MIT. It adds 77 tokens to every session and 1,226 once invoked, about $0.0003 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
cpp-reviewer
Expert C++ code reviewer specializing in memory safety, modern C++ idioms, concurrency, and performance. Use for all C++ code changes. MUST BE USED for C++ projects.
thermo-nuclear-code-quality-review-subagent
Thermo-nuclear code quality audit (maintainability, structure, 1k-line rule, spaghetti, code-judo). Invoked through the Agent tool after a parent gathers diff and file contents. Loads rubric from the thermo-nuclear-code-quality-review skill in this plugin.
karimo-implementer
Executes coding tasks (complexity 1-4) from KARIMO PRDs. Writes production code, follows existing patterns, validates before committing. Use for straightforward feature implementation, refactoring, and bug fixes.
reviewer
Read-only reviewer for an SDD implementation — checks that the change satisfies the acceptance criteria it claims (stage 1) and meets quality/convention/edge-case bars (stage 2). Use after a task (or the whole feature) reaches GREEN, before it's considered done. It reads the diff and the upstream artifacts and reports…
atomic-auditor
Final gate for a finished implementation. Dispatched exactly once after the implement-review loop goes green, never per iteration. Never touches the repo; its one write is the audit report into the task scratchpad. Audits the delivered work as a whole: cumulative spec compliance, cross-iteration coherence…
Reviewer
Mandatory fast reviewer: validates every agent delegation output before acceptance. Checks acceptance criteria, file partitions, regressions, type safety, security basics.