@papi-ai/server 0.7.40 → 0.7.42
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/design-assets/agents/frontend-design-engineer.md +141 -0
- package/design-assets/hooks/frontend-design-guard.sh +52 -0
- package/design-assets/skills/design-critique/SKILL.md +173 -0
- package/dist/backfill-cycle-metrics.js +116 -3
- package/dist/index.js +1175 -511
- package/package.json +3 -2
- package/skills/papi-cycle/papi-advanced/SKILL.md +5 -10
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@papi-ai/server",
|
|
3
|
-
"version": "0.7.
|
|
3
|
+
"version": "0.7.42",
|
|
4
4
|
"description": "PAPI MCP server — AI-powered sprint planning, build execution, and strategy review for software projects",
|
|
5
5
|
"license": "Elastic-2.0",
|
|
6
6
|
"mcpName": "io.github.cathalos92/papi",
|
|
@@ -16,11 +16,12 @@
|
|
|
16
16
|
"files": [
|
|
17
17
|
"dist",
|
|
18
18
|
"skills",
|
|
19
|
+
"design-assets",
|
|
19
20
|
"README.md"
|
|
20
21
|
],
|
|
21
22
|
"scripts": {
|
|
22
23
|
"build": "npm run build:skills && tsup",
|
|
23
|
-
"build:skills": "
|
|
24
|
+
"build:skills": "node scripts/export-skills.mjs --out skills/papi-cycle",
|
|
24
25
|
"build:prompts": "tsup src/prompts.ts --format esm",
|
|
25
26
|
"start": "node dist/index.js",
|
|
26
27
|
"backfill-cycle-metrics": "node dist/backfill-cycle-metrics.js",
|
|
@@ -7,18 +7,13 @@ description: Invoke for cross-project patterns, dogfood-derived workflows, or ad
|
|
|
7
7
|
|
|
8
8
|
## When to Start a New Conversation
|
|
9
9
|
|
|
10
|
-
|
|
11
|
-
- **After a release** — cycle is done, context is heavy. New window orients in seconds via `orient`.
|
|
12
|
-
- **After 3+ tasks built** — accumulated file reads, diffs, and discussions bloat context. Quality degrades.
|
|
13
|
-
- **Switching modes** — going from building to planning, or from strategy review to building. Each mode benefits from clean context.
|
|
14
|
-
- **After context compression fires** — if you notice earlier messages are missing, the window is getting stale. Open fresh.
|
|
10
|
+
Your AI host manages its own context — it compacts automatically as a conversation grows, so you rarely need to restart. **Stay in the same conversation by default.** Keep building, planning, and reviewing in one place; `orient` re-grounds you whenever you need it.
|
|
15
11
|
|
|
16
|
-
|
|
17
|
-
-
|
|
18
|
-
-
|
|
19
|
-
- Having a strategic discussion that informs the next action
|
|
12
|
+
Only consider a fresh conversation when:
|
|
13
|
+
- **Context pressure is real** — your host signals it's compacting, or you notice earlier messages have dropped. Let that be the trigger, not a timer or a task count.
|
|
14
|
+
- **You deliberately change modes** — e.g. you've just released a cycle and are sitting down to plan the next one, where a clean slate genuinely helps.
|
|
20
15
|
|
|
21
|
-
|
|
16
|
+
Do NOT restart based on elapsed time or number of tasks built. A fresh conversation costs a full re-orient — only pay it when context pressure is actually there.
|
|
22
17
|
|
|
23
18
|
## Advanced Patterns
|
|
24
19
|
|