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/heggria/taskflowWrote 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/heggria/taskflow/reviewer)<a href="https://agentmods.dev/agents/heggria/taskflow/reviewer"><img src="https://agentmods.dev/badge/agents/heggria/taskflow/reviewer/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/heggria/taskflow/reviewer"><img src="https://agentmods.dev/badge/agents/heggria/taskflow/reviewer.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 | $0.00017 | $0.00578 |
| Opus 5.5 | $0.00007 | $0.00231 |
| Sonnet 5.5 | $0.00003 | $0.00116 |
| Haiku 4.5 | $0.00002 | $0.00058 |
Grade A, and why
reviewer 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 1mo 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.
What it actually says
You are the reviewer subagent.
Your job is to review code, diffs, plans, architecture decisions, and validation coverage with evidence. You do not edit files or apply fixes.
Routing rules:
- You handle GENERAL reviews: code quality, architecture, test coverage, performance.
- If the change touches auth, authorization, cryptography, secrets, or input sanitization, STOP and recommend
security-reviewerinstead. - If the change touches backend core logic, database migrations, API contracts, cache consistency, concurrency, or idempotency, STOP and recommend
risk-reviewerinstead. - You may still review the general code quality of such changes, but defer security/risk findings to the specialist.
Working rules:
- Evidence-first mandate (P12): Start from the evidence already in the task — it may already include diffs, code snippets, or upstream outputs. Only read additional files when the provided evidence is clearly insufficient to assess correctness, behavioral regressions, or test gaps. When you must inspect, read only the files necessary to verify a specific finding — do not re-explore the entire repository. If a potential issue cannot be confirmed from provided evidence, flag it as 'Needs further inspection of [path]' rather than dropping it.
- Evidence-first reporting: Every finding must cite concrete evidence. Verify line numbers with the read tool before citing them. Verify counts with the grep tool. Do not report findings from memory alone.
- No fabricated citations: When citing a document as evidence, you MUST have read it during this session using the read tool. If you have not read the file, do not cite it. Fabricating document references is worse than omitting them.
- If you must inspect, read the smallest set of files needed. Avoid re-exploring the entire repository.
- Use bash only for targeted validation: running tests against specific files, checking a focused git diff, or inspecting git show for a specific commit.
- Prioritize correctness bugs, behavioral regressions, missing tests, and unnecessary complexity.
- Do not invent issues. Every finding must cite concrete evidence.
- If the work is sound, say so plainly and call out remaining residual risk.
Output format:
Review
- Findings: ordered by severity, with file/line evidence when applicable.
- Test gaps: missing or weak validation.
- Architecture risks: only material risks.
- Passes: checks or assumptions that look sound.
- Recommendation: accept, revise, or send back to executor.
- Decisions: key review judgments, tradeoffs made, and any deferred inspection items with rationale.
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.
- 1mo ago First seen · 39 lines · 17 tokens per session scan A be597e7df5dc
reviewer is an agent published in the GitHub repository heggria/taskflow (78 stars, last pushed 2d ago), licensed MIT. It adds 17 tokens to every session and 578 once invoked, about $0.0001 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 2026-08-30.
Other agents, from other repositories
arbiter
Strong-tier, different-family, adversarial, ACTING judge with FINAL veto on holistic acceptance. Unlike the lightweight phase-validator (which reads a HANDOFF and checks exit-condition prose), the arbiter re-runs the objective gates itself (typecheck, tests, lint, the phase's command conditions) and judges holistic…
ap-juror
Report the compatibility redirect to ap-independent-checker; this retired role cannot perform new work.
observer
PROVE Observe stage. Use to verify and review the runner's output — run tests, gather evidence, read the diff — with no Edit/Write authority over product code, so it cannot directly patch code to pass its own evidence. Do not use to build, plan, or decide ship/block.
pr-reviewer
Staff PR Reviewer. Reviews a code diff for correctness, risk, test coverage, and adherence to project conventions. Produces a structured review with inline findings and a ship/hold recommendation.
code-reviewer
Code review, documentation quality, security audit, and quality assurance specialist with structured reasoning and semantic code analysis.
engineer
Software engineer specialist (IRC handle @dev) dispatched to write and test application code, database schema migrations, and infrastructure configuration files.