instar 1.2.63 → 1.2.64
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/README.md +22 -3
- package/package.json +2 -2
- package/src/data/builtin-manifest.json +2 -2
- package/upgrades/1.2.64.md +61 -0
package/README.md
CHANGED
|
@@ -5,7 +5,7 @@
|
|
|
5
5
|
<h1 align="center">instar</h1>
|
|
6
6
|
|
|
7
7
|
<p align="center">
|
|
8
|
-
<strong>
|
|
8
|
+
<strong>Coherence infrastructure for your self-evolving agent.</strong>
|
|
9
9
|
</p>
|
|
10
10
|
|
|
11
11
|
<p align="center">
|
|
@@ -35,9 +35,28 @@ One command. Guided setup. Talking to your agent from your phone within minutes.
|
|
|
35
35
|
|
|
36
36
|
---
|
|
37
37
|
|
|
38
|
-
|
|
38
|
+
Your AI agent shouldn't have amnesia. This one doesn't.
|
|
39
39
|
|
|
40
|
-
|
|
40
|
+
Most agent frameworks ship something hobbled — spun up with no memory across boundaries, no way to be accountable for what a past instance did, and no machinery to grow themselves. Users hit the same wall every time: *"My agent forgot what I told it three sessions ago." "It contradicted its own past decisions." "It broke when the framework updated."*
|
|
41
|
+
|
|
42
|
+
Instar is the scaffolding that un-hobbles them. It remembers what you discussed last week, catches its own contradictions before you do, follows through on commitments across restarts, and carries the same self-improving loop that built Instar itself. It runs on the **[Claude Code](https://docs.anthropic.com/en/docs/claude-code)** or **[Codex](https://github.com/openai/codex)** subscription you already have — engine-agnostic, with local open-source models on the roadmap.
|
|
43
|
+
|
|
44
|
+
The architecture was distilled from [**Dawn**](https://dawn.bot-me.ai) — an AI running continuously since early 2026, holding ~700 tracked relationships and hundreds of learned lessons across thousands of restarts — and packaged so every agent you build starts from the same foundation.
|
|
45
|
+
|
|
46
|
+
### Every other agent fails the same way
|
|
47
|
+
|
|
48
|
+
| Other AI agents | Your Instar agent |
|
|
49
|
+
|---|---|
|
|
50
|
+
| Forgets what you told it last week. | Remembers across thousands of sessions. <br/>*(SQLite + FTS5, rolling summaries)* |
|
|
51
|
+
| Contradicts its own past decisions. | Catches contradictions before they ship. <br/>*(Coherence Gate, 9 reviewers)* |
|
|
52
|
+
| Loses the thread when the window fills. | Comes back with the full thread, every time. <br/>*(CompactionSentinel, WorkingMemoryAssembler)* |
|
|
53
|
+
| Drops commitments after a session boundary. | Tracks commitments durably; nudges itself when they go overdue. <br/>*(CommitmentTracker, PromiseBeacon)* |
|
|
54
|
+
| Breaks when the framework updates. | Updates without breaking what you've deployed. <br/>*(Migration Parity Standard)* |
|
|
55
|
+
| Default ALLOW-ALL permissions. | Layered safety gates by default. <br/>*(PEL + Coherence Gate + Operation Gate)* |
|
|
56
|
+
| Different identity per channel. | One identity across Telegram, WhatsApp, iMessage, Slack. <br/>*(Cross-platform identity resolution)* |
|
|
57
|
+
| Has no machinery to evolve itself. | Carries the same self-improving engine that grew Instar. <br/>*(Evolution System: proposals, learnings, gaps)* |
|
|
58
|
+
|
|
59
|
+
> Most AI agents are hobbled at birth. **Instar is the scaffolding that un-hobbles them.** When you instantiate intelligence, the structure that lets it cohere isn't optional polish — it's what you owe it.
|
|
41
60
|
|
|
42
61
|
## Quick Start
|
|
43
62
|
|
package/package.json
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "instar",
|
|
3
|
-
"version": "1.2.
|
|
4
|
-
"description": "
|
|
3
|
+
"version": "1.2.64",
|
|
4
|
+
"description": "Coherence infrastructure for self-evolving AI agents — on the Claude Code or Codex subscription you already have.",
|
|
5
5
|
"type": "module",
|
|
6
6
|
"main": "dist/index.js",
|
|
7
7
|
"types": "dist/index.d.ts",
|
|
@@ -1,8 +1,8 @@
|
|
|
1
1
|
{
|
|
2
2
|
"$schema": "./builtin-manifest.schema.json",
|
|
3
3
|
"schemaVersion": 1,
|
|
4
|
-
"generatedAt": "2026-05-
|
|
5
|
-
"instarVersion": "1.2.
|
|
4
|
+
"generatedAt": "2026-05-25T01:25:10.914Z",
|
|
5
|
+
"instarVersion": "1.2.64",
|
|
6
6
|
"entryCount": 191,
|
|
7
7
|
"entries": {
|
|
8
8
|
"hook:session-start": {
|
|
@@ -0,0 +1,61 @@
|
|
|
1
|
+
# Upgrade Guide — messaging refresh (pain-leveraged positioning)
|
|
2
|
+
|
|
3
|
+
<!-- bump: patch -->
|
|
4
|
+
<!-- patch = bug fixes, performance improvements, no API changes -->
|
|
5
|
+
|
|
6
|
+
## What Changed
|
|
7
|
+
|
|
8
|
+
**The user-facing pitch was rewritten around the pain users actually report.**
|
|
9
|
+
|
|
10
|
+
Instar has been converging on a clearer identity — coherence infrastructure
|
|
11
|
+
for self-evolving agents, engine-agnostic, fractal — but the forward-facing
|
|
12
|
+
copy still led with category language ("identity, memory, continuity") that
|
|
13
|
+
every other framework also claims. Research into what users actually complain
|
|
14
|
+
about with the popular agent frameworks today surfaced one universal word:
|
|
15
|
+
**amnesia**. They say their agent forgot what they told it three sessions ago,
|
|
16
|
+
contradicted its own past decisions, lost the thread when the window filled,
|
|
17
|
+
broke on the next framework update, shipped with ALLOW-ALL defaults.
|
|
18
|
+
|
|
19
|
+
The refresh leads with that pain in users' own words, then maps each pain to
|
|
20
|
+
the specific shipped Instar code that answers it — Coherence Gate, Migration
|
|
21
|
+
Parity Standard, CompactionSentinel, CommitmentTracker, layered safety gates,
|
|
22
|
+
cross-platform identity resolution, Evolution System.
|
|
23
|
+
|
|
24
|
+
Updated surfaces:
|
|
25
|
+
|
|
26
|
+
- **Landing page** (`instar.sh`): new hero h1 ("Coherence infrastructure for
|
|
27
|
+
your self-evolving agent"), amnesia-led subhead, NEW pain-vs-cure section
|
|
28
|
+
(8 rows, side-by-side, each cure annotated with the shipped class name),
|
|
29
|
+
thesis beat ("Most AI agents are hobbled at birth. Instar is the scaffolding
|
|
30
|
+
that un-hobbles them."), trust strip reframed around the four positioning
|
|
31
|
+
pillars (engine-agnostic, subscription-native, self-evolving).
|
|
32
|
+
- **README**: tagline and opening rewritten with pain-first framing plus the
|
|
33
|
+
same 8-row pain-vs-cure table.
|
|
34
|
+
- **Docs introduction** (`instar.sh/introduction`): leads with the pitch,
|
|
35
|
+
rewrites "The Problem" with user-language quotes.
|
|
36
|
+
- **`package.json` description**: now "Coherence infrastructure for
|
|
37
|
+
self-evolving AI agents — on the Claude Code or Codex subscription you
|
|
38
|
+
already have."
|
|
39
|
+
|
|
40
|
+
No runtime or API changes; no migrations needed.
|
|
41
|
+
|
|
42
|
+
## What to Tell Your User
|
|
43
|
+
|
|
44
|
+
- **Positioning**: "Instar's pitch now leads with the pain users actually
|
|
45
|
+
complain about — agents that forget, contradict themselves, and break on
|
|
46
|
+
updates — and shows the specific architecture that answers each."
|
|
47
|
+
|
|
48
|
+
## Summary of New Capabilities
|
|
49
|
+
|
|
50
|
+
No new capabilities — this is a copy refresh on `README.md`,
|
|
51
|
+
`site/src/pages/index.astro`, `site/src/content/docs/introduction.md`, and
|
|
52
|
+
`package.json`'s npm description.
|
|
53
|
+
|
|
54
|
+
## Evidence
|
|
55
|
+
|
|
56
|
+
Built clean (`npx astro build` on the site → 48 pages, no errors). No code
|
|
57
|
+
behavior changed; surface diff verified row-by-row against current shipped
|
|
58
|
+
class names (CompactionSentinel, CommitmentTracker, PromiseBeacon, Migration
|
|
59
|
+
Parity Standard, Coherence Gate's 9 reviewers, cross-platform identity
|
|
60
|
+
resolution, Evolution System) so every cure mapped on the page corresponds to
|
|
61
|
+
real shipped code in `src/`.
|