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/tajd/projektor/workflow-specgit clone --depth 1 https://github.com/TAJD/projektorWhat 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 | $0.00020 | $0.01207 |
| Opus 5 | $0.00010 | $0.00603 |
| Sonnet 5 | $0.00004 | $0.00241 |
| Haiku 4.5 | $0.00002 | $0.00121 |
Grade A, and why
workflow-spec 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 2d 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 — 102 lines — stays where its author put it; the contents beside it link to each section on GitHub.
This page is the single home for projektor's agentic workflow rules. Every other
surface — the MCP server's initialize instructions, AGENTS.md, spawn prompts, skills —
points here instead of restating these rules. If you're reading a copy of this text
somewhere else, that copy is stale; this page wins. Fetch it programmatically any time
via the get_workflow MCP tool or GET /api/workflow.
Definition of ready
An issue is ready for an agent to pick up when its body states all three of:
- Acceptance criteria — a checklist or bullet list of concrete, checkable outcomes.
- Scope — the files or components expected to change, named explicitly (not "the backend").
- Verification — the exact command(s) that prove the work is done (a test file, a lint/type-check command, a fixture run).
get_prioritized_issues filters out issues that don't meet this bar by default. Pass
includeNotReady: true to see them anyway, tagged with needsGrooming: true and the
specific criteria missing — useful for a human doing backlog grooming, not for an agent
picking up autonomous work.
State machine
| State | Meaning |
|---|---|
| Backlog | Not yet triaged into a workable slice. |
| Ready | Triaged (status: todo) and passes the definition of ready above. |
| Claimed | An agent (or human) holds a live lease via claim_issue; work is in progress. |
| In Review | Work is done from the implementer's side; a structured completion report is attached. |
| Done | Reviewed and accepted. |
| Cancelled | Won't do. |
Human gates
- Ready → Claimed: may be fully autonomous. Any live agent session can call
claim_issuewithout a human in the loop, subject to the WIP limit below. - In Review → Done: may also be fully autonomous. An agent session can
close its own (or any) issue to
donedirectly — there's no pre-close block waiting on a human. An agent-worked issue still needs a completion report before it can close (see below). Instead of gating the transition, projektor classifies the report's evidence and flags weak closures for audit after the fact — see below.
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.
- 2d ago First seen · 102 lines · 20 tokens per session scan A cb609a699452
workflow-spec is an agent published in the GitHub repository TAJD/projektor (39 stars, last pushed 7d ago), licensed MIT. It adds 20 tokens to every session and 1,207 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
discovery
How an agent finds these surfaces without being told: the API catalog, DNS-based discovery, an agent-skills index, and the places people actually look.
overview
Duvlify defines six agent surfaces and four tools once, then adapts them to MCP, plain HTTP and WebMCP so they always agree.
yegge
Primary session agent for this project. Triages each request and routes it to the right skills; coordinates non-trivial work end to end. All user requests come through it.
linear-issue-creator
Create Linear issues from user descriptions. NOT for URLs — use linear-url-to-issues for that. Use when: "create issue", "new ticket", "log bug", "add task", "file issue", "report bug", "make ticket", "make issue", "add to linear", "create task", "new issue", "make new issue", "make new ticket", "make new linear…
linear-url-to-issues
Extract actionable Linear issues from URLs (articles, blogs, designs, docs). Parallel-safe — dispatch one agent per URL when processing multiple. NOT for manual creation — use linear-issue-creator. Use when: "create issues from URL", "turn this article into tasks", "implement this design", "make issues from blog…
triage-labels
The skills speak in terms of five canonical triage roles. This file maps those roles to the actual label strings used in this repo's issue tracker.