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 KirKruglov/claude-project-manager --skill generate-project-plangit clone --depth 1 https://github.com/KirKruglov/claude-project-managerWrote 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/kirkruglov/claude-project-manager/generate-project-plan)<a href="https://agentmods.dev/skills/kirkruglov/claude-project-manager/generate-project-plan"><img src="https://agentmods.dev/badge/skills/kirkruglov/claude-project-manager/generate-project-plan/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/skills/kirkruglov/claude-project-manager/generate-project-plan"><img src="https://agentmods.dev/badge/skills/kirkruglov/claude-project-manager/generate-project-plan.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 | $0.00124 | $0.03814 |
| Opus 5 | $0.00062 | $0.01907 |
| Sonnet 5 | $0.00025 | $0.00763 |
| Haiku 4.5 | $0.00012 | $0.00381 |
Grade A, and why
generate-project-plan 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 10d 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.
This is a copy
95% identical to generate-project-plan — 16 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 261 lines — stays where its author put it; the contents beside it link to each section on GitHub.
generate-project-plan
Language Detection
Determine the language of the user's request:
- If the request is in Russian → use templates with
-rusuffix - Otherwise → use templates with
-ensuffix
All output (headings, labels, comments, instructions) must match the detected language.
Triggers
Russian: «сформируй план проекта», «подготовь план», «составь project plan», «нужен план проекта» English: "generate project plan", "create project plan", "build plan", "prepare project plan"
1. Inputs
| Data | Required | Source | Notes |
|---|---|---|---|
| Project charter (project-charter.md) | yes | knowledge or chat | Approved charter. Extract: goal, scope, dates, budget (if present), team, constraints, top-3 risks |
| Duration estimates | no | chat or knowledge | Work package estimates from the user. If absent — agent estimates and marks as assumption |
| Team availability data | no | chat or knowledge | Load by period, constraints (vacations, part-time). If absent — agent distributes evenly, marks as assumption |
| Additional inputs | no | chat | Priorities, external dependencies, technical constraints not included in the charter |
If no charter is available — request it (system-prompt-draft.md §3 p.2):
A project plan requires an approved charter (project-charter.md).
Available data: [list what is available].
Charter status: [not started / draft / approved].
If the charter has not been created yet — start with generate-charter.
2. Execution Algorithm
- Validate inputs. Check for an approved charter. Not found → request per §1 template. Do NOT generate a plan without a charter — this breaks the dependency chain. Found → step 2.
- Extract data from the charter. Break down by category: goal, scope (included/excluded), dates (start/end), budget (if specified), team, constraints, risks. Record gaps.
- Check data for conflicts. If charter dates conflict with user-provided duration estimates — do NOT resolve independently. List discrepancies, ask which takes precedence. Exception: if the user explicitly clarifies in chat — use the clarification.
- Decompose scope into work packages (WP). Each deliverable from the charter's "Included" section → one or more WPs. Granularity: a WP is a group of related tasks with a single output. Do NOT decompose to individual tasks — that is Jira level. Assign each WP an ID (WP-001, WP-002, etc.). Group WPs by project phase. After decomposition, count total WPs: if >20 — consolidate adjacent packages within the same phase. WP count must be ≤ 20 for MVP. If consolidation would lose meaning — record as assumption and proceed to step 5.
- Set duration for each WP.
- Use user-provided estimates if available.
- If no estimates — assess based on: scope volume, resource count, comparable projects, charter constraints. Mark each such estimate as assumption.
- Unit: project ≤ 3 months → working days (wd); project > 3 months → weeks (wk). Unit must be consistent within one plan. 1 wk = 5 wd.
- Verify: sum of critical path durations (accounting for maximum parallelism) ≤ total project duration from charter. If it doesn't fit — record as issue, propose options (parallel execution / scope reduction / resource increase / date shift). Do NOT generate WP dates beyond the charter end_date. If unavoidable — mark affected WPs with
[⚠ exceeds deadline]and require a decision from the user before finalizing.
- Identify dependencies between WPs. For each related WP pair: dependency type (FS / SS / FF). FS is the default if the relationship is not explicit. Verify no circular dependencies.
- Define milestones. A milestone is a checkpoint marking the achievement of a significant result. A milestone is NOT a task — it has zero duration. It answers "what must be done by this point", not "what needs to be done". What counts as a milestone: phase completion (all WPs in phase closed); delivery of a key deliverable to the customer; external event the project depends on (approval, procurement, integration); go/no-go decision point. What does NOT count as a milestone: completion of a routine WP that is not a deliverable; internal intermediate results without external review; recurring events (weekly status, sprint reviews). Milestones from the charter (section 3) — must be included. ID: M1, M2, etc. Minimum: 3 milestones (start, mid, finish). Maximum: one per phase + key deliverables.
- Fill the resource map.
- Use availability data if provided.
- If not — take team from charter (section 5), distribute across WPs by role, treat load as uniform. Mark as assumption.
- Period columns — actual project months (not fixed).
- Read the template
project-plan-ru.md(orproject-plan-en.md) from project knowledge. - Fill the template per the rules in section 3.
- Validate the output using the checklist (section 5).
- Present the output in chat. Format (system-prompt-draft.md §3 p.4):
- What was done: "Project plan generated based on [list sources]."
- Assumptions: list all (duration estimates, resource allocation, etc.).
- Issues: if found (schedule overrun, insufficient resources).
- What is required: "Approve, provide comments, or reject."
- Do not include system paths. Use file name only: "Plan saved as project-plan.md."
- Wait for user response (system-prompt-draft.md §3 p.5).
- Approval → generate final version.
- Comments → revise, show updated version. After 3rd iteration — ask: "Continue revising or finalize the current version?"
- After approval:
- Propose text for the user to insert into log.md and project-state.md (system-prompt-draft.md §11).
- Notify: "Project plan approved. The following tasks are now available (can run in parallel): generate-risk-register (detailed risk register) and generate-comm-plan (communications plan). Which one to start?" (system-prompt-draft.md §7).
What ships with it
2 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.
- 10d ago First seen · 261 lines · 124 tokens per session scan A c7493524de4d
generate-project-plan is a skill published in the GitHub repository KirKruglov/claude-project-manager (6 stars, last pushed 4mo ago), licensed MIT. It adds 124 tokens to every session and 3,814 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it A with 0 findings. It is 95% identical to generate-project-plan, differing in 16 lines, and is treated as a copy.
Other skills, from other repositories
litigation-workflow
Civil litigation workflow expertise — case management, deposition prep, evidence chain, and Federal Rules of Evidence.
construction-management
Estimating, scheduling, building codes, safety protocols, permitting, and project controls.
stakeholder-communication
Owner reports, subcontractor coordination, RFI responses, and meeting documentation.
run-team-retrospective
Use when a team needs to reflect on a completed period, project, or significant event — to extract what worked, what didn't, and what specific changes to make before the next cycle, rather than repeating the same patterns.
project-manager
This skill has been upgraded with agentic AI capabilities, OKR/KPI integration patterns, and async-first workflows based on 2026 PM best practices research.
github-project-management
Comprehensive GitHub project management with swarm-coordinated issue tracking, project board automation, and sprint planning.