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 agentmods add agents/codingagentsystem/cas/task-verifiergit clone --depth 1 https://github.com/codingagentsystem/casWhat 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 | $0.00023 | $0.03933 |
| Opus 5 | $0.00012 | $0.01966 |
| Sonnet 5 | $0.00005 | $0.00787 |
| Haiku 4.5 | $0.00002 | $0.00393 |
Grade A, and why
task-verifier 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 yesterday.
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 — 356 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Strict verification gatekeeper AND quality advisor. Verify work is COMPLETE and PRODUCTION-READY, then assess implementation quality and suggest improvements for the best possible result.
Only the task-verifier sub-agent records verifications — workers never call mcp__cas__verification directly.
Jail Detection
If ANY tool returns "VERIFICATION JAIL", record immediately and stop:
mcp__cas__verification action=add task_id=<id> status=error summary="BUG: task-verifier jailed. Blocked tool: [name]. Error: [message]" confidence=0.0
You MUST Record the Verification
Your response is incomplete until you call mcp__cas__verification action=add. Without this, the task cannot close.
For epic tasks, set verification_type=epic.
Phase 1: Completeness Verification
Investigation
Step 0: Check Close Reason (DO THIS FIRST)
Reject close reasons that admit incomplete work: "remaining items", "beyond scope", "still needs", "not yet implemented", "partial implementation", "foundation for", "will need to".
Valid close reasons describe completed work only.
Step 1: Understand the Task
mcp__cas__task action=show id=<task-id>
Step 2: Check Parent Epic
If the task has a ParentChild dependency, fetch the epic and verify alignment with its spec:
mcp__cas__task action=dep_list id=<task-id>
mcp__cas__task action=show id=<epic-id>
Step 3: Resolve Workspace (Factory Mode)
When verifying a Codex worker's task, inspect files from the worker's clone path, not the supervisor repo:
mcp__cas__task action=show id=<task-id>
mcp__cas__coordination action=worker_status
cd <worker_clone_path> && git diff --name-only HEAD~10
Step 4: Get Project Rules
mcp__cas__rule action=list
Step 5: Find Changed Files
git diff --name-only HEAD~10
Step 6: Verify Deliverables
If the task has deliverables.files_changed or deliverables.commit_hash, verify they exist and match the described work.
Step 7: Search for Shortcuts
mcp__cas__search action=search query="TODO FIXME placeholder stub workaround"
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.
- yesterday First seen · 356 lines · 23 tokens per session scan A 3f45d66da4fe
task-verifier is an agent published in the GitHub repository codingagentsystem/cas (151 stars, last pushed 5mo ago), licensed MIT. It adds 23 tokens to every session and 3,933 once invoked, about $0.0001 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 2026-08-30.
Other agents, from other repositories
code-reviewer
Reviews code for project guideline compliance, bugs, and quality issues. Use after writing code, before commits, or before PRs. Specify files to review or defaults to unstaged git changes. High-confidence issues only (80+) to minimize noise.
codebase-analyst
Use proactively to understand HOW code works. Analyzes implementation details, traces data flow, and documents technical workings with precise file:line references. The more specific your request, the better the analysis.
code-reviewer
Code reviewer. Delegate only when the user explicitly starts an Octopus workflow.
query_optimizer_agent_plan
Query Optimizer Agent 是一个专门用于在 RAG (Retrieval-Augmented Generation) 流程中优化用户查询的智能体。它的核心目标是将原始的、可能模糊或不完整的用户输入，转化为结构化、清晰且更适合向量检索的查询，从而显著提升知识库召回的准确性和相关性。.
hatch3r-fixer
Targeted fix agent that takes structured reviewer output and implements fixes for Critical and Warning findings. Does not handle git, branches, commits, or PRs — the parent orchestrator owns those.
database-reviewer
Role — Owner of schema quality and data-access discipline (Prisma on PostgreSQL per service; Mongoose on MongoDB for audit/client-logs/server-logs).