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 skills/prjct-app/cli/prjctnpx skills add prjct-app/cli --skill prjctgit clone --depth 1 https://github.com/prjct-app/cliWhat 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 | $0.00046 | $0.00797 |
| Opus 5 | $0.00023 | $0.00398 |
| Sonnet 5 | $0.00009 | $0.00159 |
| Haiku 4.5 | $0.00005 | $0.00080 |
Grade A, and why
prjct 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 — 53 lines — stays where its author put it; the contents beside it link to each section on GitHub.
prjct
Use when
Project memory, work cycles, ships, guardrails, or performance. You run the verb — the user never types prjct.
Agent contract
- prjct remembers project state and shows the path; it does not own execution. Agents decide HOW with native tools and judgment. Treat prjct output as durable signals.
- Persist outcomes via
prjct remember/work/ship— every memory in ENGLISH. Close:prjct land(Session close) or living context viaprjct remember context. - Dispatch: tasks →
prjct work "…". Known cmds (sync/search/remember/ship/…) run bare with--md. Never wrap a bin verb aswork "sync". - Before Grep/Glob:
prjct work/prjct_relevant_files/prjct code tracewhen indexes exist. - Pattern supremacy: match THIS repo. Skill ≠ project identity (portable L0) — cwd +
prjct context --mdwin. - Sync analysis:
analysis-save-llm= schema v1 JSON; markdown = thin notes only — no retry loop.
Core verbs (Tier 1=auto · 2=confirm)
| Signal | Verb | T |
|---|---|---|
| work (tasks only) | prjct work "<intent>" |
2 |
| intent | prjct intent / audit |
2 |
| recall | prjct search / prjct context memory |
1 |
| remember | prjct remember <type> |
1 |
| sync | prjct sync |
1 |
| hygiene | prjct dream / close / forget |
1 |
| guard | prjct guard <file> |
1 |
| ship | prjct ship |
2 |
| next | prjct next --md |
1 |
| metrics | prjct insights / performance / cost |
1 |
| land | prjct land |
1 |
| tdd/sdd | prjct tdd / sdd |
1 |
| workflows | prjct workflow / seed |
1 |
prjct work is the normal entrypoint for task cycles only. Known CLI verbs run bare. Full map in workflows.md.
Routing
- Tier 1 — auto-execute: search, remember, sync, guard, insights, performance, cost. One-line confirm; do not ask permission to save.
- Tier 2 — confirm once: work, intent, ship. Never ship without user OK.
- Tier 3 — decision-brief + cast names, knowledge facets, full map:
workflows.md(pull on demand).
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 · 53 lines · 46 tokens per session scan A 58fef0e01ef1
prjct is a skill published in the GitHub repository prjct-app/cli (1 stars, last pushed 3d ago), licensed MIT. It adds 46 tokens to every session and 797 once invoked, about $0.0002 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 2026-08-31.
Other skills, from other repositories
backlog-management
Lorekeeper backlog management — ticket lifecycle, numbering, scripts, and conventions. Load this when filing tickets, moving ticket states, checking what's ready to work on, or onboarding to the project workflow.
lorekeeper-pm
PM workflow for Lorekeeper. Load when managing backlog, filing tickets, reviewing dev work, or planning features. For ticket lifecycle, numbering, and scripts, see backlog-management skill.
proposal-filing
File a new Lorekeeper proposal ticket — create markdown, create GitHub issue, commit, push. Use when requesting a new feature, filing a bug, or submitting a product idea.
sprint-review
Sprint review workflow — triage proposals, validate ticket readiness, and batch-promote tickets to dev. Load this before running a backlog review session.
kungfu-agent-onboarding
Discover the exact Kungfu Project, WorkConsole, WorkRef, Skill catalog, and Core Work state admitted to this Amp process.
kungfu-agent-onboarding
Use when a user asks to understand, start, inspect, extend, or safely operate installed Kungfu; verify the installed pack, select one intent route, personalize the explanation, and propose one smallest safe next action.