Commit 6ba36cf
authored
![dependabot[bot]](https://avatars.githubusercontent.com/in/29110?v=4&size=40){kind=avatar}
chore(deps): bump @mariozechner/pi-ai from 0.67.68 to 0.70.2 (#232)
Bumps
[@mariozechner/pi-ai](https://github.com/badlogic/pi-mono/tree/HEAD/packages/ai)
from 0.67.68 to 0.70.2.
<details>
<summary>Release notes</summary>
<p><em>Sourced from <a
href="https://github.com/badlogic/pi-mono/releases"><code>@mariozechner/pi-ai</code>'s
releases</a>.</em></p>
<blockquote>
<h2>v0.70.2</h2>
<h3>Fixed</h3>
<ul>
<li>Fixed provider retry/timeout forwarding to omit undefined provider
request controls, avoiding downstream SDK validation errors such as
<code>timeout must be an integer</code> when
<code>retry.provider.timeoutMs</code> is not configured (<a
href="https://redirect.github.com/badlogic/pi-mono/issues/3627">#3627</a>)</li>
</ul>
<h2>v0.70.1</h2>
<h3>New Features</h3>
<ul>
<li>DeepSeek provider support with V4 Flash/Pro models and
<code>DEEPSEEK_API_KEY</code> authentication. See <a
href="https://github.com/badlogic/pi-mono/tree/HEAD/packages/ai/blob/HEAD/README.md#providers--models">https://github.com/badlogic/pi-mono/tree/HEAD/packages/ai/blob/HEAD/README.md#providers--models</a>
and <a
href="https://github.com/badlogic/pi-mono/tree/HEAD/packages/ai/blob/HEAD/docs/providers.md#api-keys">https://github.com/badlogic/pi-mono/tree/HEAD/packages/ai/blob/HEAD/docs/providers.md#api-keys</a>.</li>
<li>Provider request timeout/retry controls via
<code>retry.provider.{timeoutMs,maxRetries,maxRetryDelayMs}</code>,
useful for long-running local inference and provider SDK retry behavior.
See <a
href="https://github.com/badlogic/pi-mono/tree/HEAD/packages/ai/blob/HEAD/docs/settings.md#retry">https://github.com/badlogic/pi-mono/tree/HEAD/packages/ai/blob/HEAD/docs/settings.md#retry</a>.
(<a
href="https://redirect.github.com/badlogic/pi-mono/issues/3627">#3627</a>)</li>
</ul>
<h3>Added</h3>
<ul>
<li>Added DeepSeek to built-in provider setup, default model resolution,
and provider documentation.</li>
</ul>
<h3>Fixed</h3>
<ul>
<li>Fixed <code>/copy</code> to avoid unbounded OSC 52 writes and
clipboard races that could break terminal rendering or panic the native
clipboard addon (<a
href="https://redirect.github.com/badlogic/pi-mono/issues/3639">#3639</a>)</li>
<li>Fixed extension flag docs to show <code>pi.getFlag()</code> using
registered flag names without the CLI <code>--</code> prefix (<a
href="https://redirect.github.com/badlogic/pi-mono/issues/3614">#3614</a>)</li>
<li>Fixed provider retry/timeout settings wiring by adding
<code>retry.provider.{timeoutMs,maxRetries,maxRetryDelayMs}</code>,
migrating legacy <code>retry.maxDelayMs</code>, and forwarding provider
controls into <code>streamSimple</code> request options (<a
href="https://redirect.github.com/badlogic/pi-mono/issues/3627">#3627</a>)</li>
<li>Fixed Windows git package installs to bypass <code>cmd.exe</code>
for native git commands, so install paths containing spaces no longer
break <code>pi install git:...</code> with <code>fatal: Too many
arguments</code> (<a
href="https://redirect.github.com/badlogic/pi-mono/issues/3642">#3642</a>)</li>
<li>Fixed DeepSeek V4 session replay 400 errors by sending
DeepSeek-compatible thinking controls and replayed assistant
<code>reasoning_content</code> fields (<a
href="https://redirect.github.com/badlogic/pi-mono/issues/3636">#3636</a>)</li>
<li>Fixed GPT-5.5 generated context window metadata to use the observed
272k limit.</li>
<li>Fixed CSI-u Ctrl+letter decoding inside bracketed paste, so pasted
modified-key escape sequences no longer become literal editor text (<a
href="https://redirect.github.com/badlogic/pi-mono/pull/3623">#3623</a>
by <a href="https://github.com/Exrun94"><code>@Exrun94</code></a>)</li>
</ul>
<h2>v0.70.0</h2>
<h3>New Features</h3>
<ul>
<li>Searchable auth provider login flow: the <code>/login</code>
provider selector now supports fuzzy search/filtering, making it faster
to find providers when many are configured. See <a
href="https://github.com/badlogic/pi-mono/tree/HEAD/packages/ai/blob/HEAD/docs/providers.md">https://github.com/badlogic/pi-mono/tree/HEAD/packages/ai/blob/HEAD/docs/providers.md</a>.
(<a
href="https://redirect.github.com/badlogic/pi-mono/pull/3572">#3572</a>
by <a
href="https://github.com/mitsuhiko"><code>@mitsuhiko</code></a>)</li>
<li>GPT-5.5 Codex support: <code>openai-codex/gpt-5.5</code> is
available as a model option, including <code>xhigh</code> reasoning
support and corrected priority-tier pricing.</li>
<li>Terminal progress indicators are now opt-in: OSC 9;4 progress
reporting during streaming/compaction is off by default and can be
toggled via <code>terminal.showTerminalProgress</code> in
<code>/settings</code> (<a
href="https://redirect.github.com/badlogic/pi-mono/issues/3588">#3588</a>)</li>
<li><code>--no-builtin-tools</code> / <code>createAgentSession({
noTools: "builtin" })</code> now correctly disables only
built-in tools while keeping extension tools active. See <a
href="https://github.com/badlogic/pi-mono/tree/HEAD/packages/ai/blob/HEAD/docs/extensions.md">https://github.com/badlogic/pi-mono/tree/HEAD/packages/ai/blob/HEAD/docs/extensions.md</a>
and <a
href="https://github.com/badlogic/pi-mono/tree/HEAD/packages/ai/blob/HEAD/README.md">https://github.com/badlogic/pi-mono/tree/HEAD/packages/ai/blob/HEAD/README.md</a>
(<a
href="https://redirect.github.com/badlogic/pi-mono/issues/3592">#3592</a>)</li>
</ul>
<h3>Breaking Changes</h3>
<ul>
<li>Disabled OSC 9;4 terminal progress indicators by default. Set
<code>terminal.showTerminalProgress</code> to <code>true</code> in
<code>/settings</code> to re-enable (<a
href="https://redirect.github.com/badlogic/pi-mono/issues/3588">#3588</a>)</li>
</ul>
<h3>Added</h3>
<ul>
<li>Added searchable auth provider login flow with fuzzy filtering in
the provider selector (<a
href="https://redirect.github.com/badlogic/pi-mono/pull/3572">#3572</a>
by <a
href="https://github.com/mitsuhiko"><code>@mitsuhiko</code></a>)</li>
<li>Added GPT-5.5 Codex model</li>
<li>Added auth source labels in <code>/login</code> so provider entries
can show when auth comes from <code>--api-key</code>, an environment
variable, or custom provider fallback without exposing secrets.</li>
</ul>
<h3>Changed</h3>
<ul>
<li>Updated default model selection across providers to current
recommended models.</li>
<li>Improved stale extension context errors after session replacement or
reload to tell extension authors to avoid captured
<code>pi</code>/command <code>ctx</code> and use
<code>withSession</code> for post-replacement work.</li>
</ul>
<!-- raw HTML omitted -->
</blockquote>
<p>... (truncated)</p>
</details>
<details>
<summary>Changelog</summary>
<p><em>Sourced from <a
href="https://github.com/badlogic/pi-mono/blob/main/packages/ai/CHANGELOG.md"><code>@mariozechner/pi-ai</code>'s
changelog</a>.</em></p>
<blockquote>
<h2>[0.70.2] - 2026-04-24</h2>
<h3>Fixed</h3>
<ul>
<li>Fixed OpenAI/Azure/Anthropic provider request option forwarding to
omit undefined <code>timeout</code>/<code>maxRetries</code>, avoiding
SDK validation errors such as <code>timeout must be an integer</code>
when provider controls are not set (<a
href="https://redirect.github.com/badlogic/pi-mono/issues/3627">#3627</a>)</li>
</ul>
<h2>[0.70.1] - 2026-04-24</h2>
<h3>Added</h3>
<ul>
<li>Added DeepSeek as a built-in OpenAI-compatible provider with V4
Flash and V4 Pro models and <code>DEEPSEEK_API_KEY</code>
authentication.</li>
</ul>
<h3>Fixed</h3>
<ul>
<li>Fixed DeepSeek V4 session replay 400 errors by adding
<code>thinkingFormat: "deepseek"</code> (sends <code>thinking:
{ type }</code> + <code>reasoning_effort</code>), a
<code>reasoningEffortMap</code>, and
<code>requiresReasoningContentOnAssistantMessages</code> compat that
injects empty <code>reasoning_content</code> on all replayed assistant
messages when reasoning is enabled (<a
href="https://redirect.github.com/badlogic/pi-mono/issues/3636">#3636</a>)</li>
<li>Fixed GPT-5.5 generated context window metadata to use the observed
272k limit.</li>
<li>Fixed provider request controls to expose <code>timeoutMs</code> and
<code>maxRetries</code> in stream options and forward them through
OpenAI/Azure/Anthropic request options, preventing unconfigurable SDK
timeout/retry defaults on long-running local inference requests (<a
href="https://redirect.github.com/badlogic/pi-mono/issues/3627">#3627</a>)</li>
</ul>
<h2>[0.70.0] - 2026-04-23</h2>
<h3>Added</h3>
<ul>
<li>Added GPT-5.5 to OpenAI Codex model generation.</li>
<li>Added <code>findEnvKeys()</code> so callers can identify configured
provider API-key environment variables without exposing credential
values while preserving <code>getEnvApiKey()</code> as the
credential-value API.</li>
</ul>
<h3>Fixed</h3>
<ul>
<li>Fixed <code>google-vertex</code> to forward custom
<code>model.baseUrl</code> values to <code>@google/genai</code>,
enabling Vertex proxy and gateway endpoints (<a
href="https://redirect.github.com/badlogic/pi-mono/issues/3619">#3619</a>)</li>
<li>Fixed OpenAI-compatible completion usage parsing to stop
double-counting reasoning tokens already included in
<code>completion_tokens</code> (<a
href="https://redirect.github.com/badlogic/pi-mono/issues/3581">#3581</a>)</li>
<li>Fixed long cache retention compatibility by adding
<code>compat.supportsLongCacheRetention</code>, allowing Anthropic
Messages and OpenAI-compatible proxies to explicitly disable
long-retention fields while enabling long retention by default when
requested (<a
href="https://redirect.github.com/badlogic/pi-mono/issues/3543">#3543</a>)</li>
<li>Fixed <code>openai-responses</code> compatibility by adding
<code>compat.sendSessionIdHeader: false</code>, allowing strict
OpenAI-compatible proxies to omit the underscore-containing
<code>session_id</code> header while still sending other
session-affinity headers (<a
href="https://redirect.github.com/badlogic/pi-mono/issues/3579">#3579</a>)</li>
<li>Fixed <code>anthropic-messages</code> tool streaming compatibility
by adding <code>compat.supportsEagerToolInputStreaming</code>, allowing
Anthropic-compatible providers to omit per-tool
<code>eager_input_streaming</code> and use the legacy fine-grained tool
streaming beta header instead (<a
href="https://redirect.github.com/badlogic/pi-mono/issues/3575">#3575</a>)</li>
<li>Fixed <code>supportsXhigh()</code> to recognize
<code>openai-codex</code> <code>gpt-5.5</code>, preserving
<code>xhigh</code> reasoning requests instead of clamping them to
<code>high</code>.</li>
<li>Fixed <code>openai-completions</code> streamed tool-call assembly to
coalesce deltas by stable tool index when OpenAI-compatible gateways
mutate tool call IDs mid-stream, preventing malformed Kimi K2.6/OpenCode
tool streams from splitting one call into multiple bogus tool calls (<a
href="https://redirect.github.com/badlogic/pi-mono/issues/3576">#3576</a>)</li>
<li>Fixed <code>packages/ai</code> E2E coverage to use currently
supported OpenAI Responses and OpenAI Codex models, and updated the
Bedrock adaptive-thinking payload expectation to match the current
<code>display: "summarized"</code> shape.</li>
<li>Fixed built-in <code>kimi-coding</code> model generation to attach
<code>User-Agent: KimiCLI/1.5</code> to all generated Kimi models,
overriding the Anthropic SDK default UA so direct Kimi Coding requests
use the provider's expected client identity (<a
href="https://redirect.github.com/badlogic/pi-mono/issues/3586">#3586</a>)</li>
<li>Fixed GPT-5.5 Codex capability handling to clamp unsupported minimal
reasoning to <code>low</code> and apply the model's 2.5x priority
service-tier pricing multiplier (<a
href="https://redirect.github.com/badlogic/pi-mono/pull/3618">#3618</a>
by <a
href="https://github.com/markusylisiurunen"><code>@markusylisiurunen</code></a>)</li>
</ul>
<h2>[0.69.0] - 2026-04-22</h2>
<h3>Breaking Changes</h3>
<ul>
<li>Migrated TypeBox support from <code>@sinclair/typebox</code> 0.34.x
plus AJV to <code>typebox</code> 1.x plus TypeBox's built-in validator
and value-conversion APIs. Tool argument validation now runs in
eval-restricted JavaScript runtimes such as Cloudflare Workers and other
environments that disallow <code>eval</code> / <code>new
Function</code>, instead of being silently skipped. Migration: install
and import from <code>typebox</code> instead of
<code>@sinclair/typebox</code>, and retest any coercion-sensitive tool
paths that serialize schemas to plain JSON because those now go through
the new TypeBox-based validation and coercion path rather than AJV (<a
href="https://redirect.github.com/badlogic/pi-mono/issues/3112">#3112</a>)</li>
</ul>
<h3>Fixed</h3>
<ul>
<li>Fixed <code>google-gemini-cli</code> built-in model discovery to
include <code>gemini-3.1-flash-lite-preview</code>, so Cloud Code Assist
model lists expose it without requiring manual <code>--model</code>
fallback selection (<a
href="https://redirect.github.com/badlogic/pi-mono/issues/3545">#3545</a>)</li>
<li>Fixed <code>transformMessages()</code> to synthesize missing
trailing tool results for transcripts that end with unresolved assistant
tool calls during direct low-level history replay (<a
href="https://redirect.github.com/badlogic/pi-mono/issues/3555">#3555</a>)</li>
</ul>
<h2>[0.68.1] - 2026-04-22</h2>
<!-- raw HTML omitted -->
</blockquote>
<p>... (truncated)</p>
</details>
<details>
<summary>Commits</summary>
<ul>
<li><a
href="https://github.com/badlogic/pi-mono/commit/48aa882b5a51e4478da1b65f98c3401151f22e3f"><code>48aa882</code></a>
Release v0.70.2</li>
<li><a
href="https://github.com/badlogic/pi-mono/commit/c1b62ae7fe70db757a349104c60c97b4a1a4eee8"><code>c1b62ae</code></a>
fix(ai,coding-agent): omit undefined provider request options closes <a
href="https://github.com/badlogic/pi-mono/tree/HEAD/packages/ai/issues/3627">#3627</a></li>
<li><a
href="https://github.com/badlogic/pi-mono/commit/90d805163631862217915f2f9b0e649aef2bddfe"><code>90d8051</code></a>
Add [Unreleased] section for next cycle</li>
<li><a
href="https://github.com/badlogic/pi-mono/commit/6b443c685ffb8c2b74912155ee143292b762cd7d"><code>6b443c6</code></a>
Release v0.70.1</li>
<li><a
href="https://github.com/badlogic/pi-mono/commit/c06750410a590fb53f01056c4f7f145d8a65c59c"><code>c067504</code></a>
fix(ai,coding-agent): expose provider timeout/retry controls closes <a
href="https://github.com/badlogic/pi-mono/tree/HEAD/packages/ai/issues/3627">#3627</a></li>
<li><a
href="https://github.com/badlogic/pi-mono/commit/1e334925254b78a74100ec87f9c255cafb4bca99"><code>1e33492</code></a>
fix(coding-agent): harden clipboard copy</li>
<li><a
href="https://github.com/badlogic/pi-mono/commit/c96c2fcd1eaf5c455082773c24d702b0fba14a3c"><code>c96c2fc</code></a>
fix(ai): correct gpt-5.5 context metadata</li>
<li><a
href="https://github.com/badlogic/pi-mono/commit/0fce64d4254ecf7097f0b4b421da6eb1aa5ca487"><code>0fce64d</code></a>
Add [Unreleased] section for next cycle</li>
<li><a
href="https://github.com/badlogic/pi-mono/commit/612be54c287a339ef092e24a7b8daa694dd82a07"><code>612be54</code></a>
Release v0.70.0</li>
<li><a
href="https://github.com/badlogic/pi-mono/commit/2926f7e8870b045d00e9fbad186ddca6833bdfe5"><code>2926f7e</code></a>
chore: update generated models</li>
<li>Additional commits viewable in <a
href="https://github.com/badlogic/pi-mono/commits/v0.70.2/packages/ai">compare
view</a></li>
</ul>
</details>
<br />
[](https://docs.github.com/en/github/managing-security-vulnerabilities/about-dependabot-security-updates#about-compatibility-scores)
Dependabot will resolve any conflicts with this PR as long as you don't
alter it yourself. You can also trigger a rebase manually by commenting
`@dependabot rebase`.
[//]: # (dependabot-automerge-start)
[//]: # (dependabot-automerge-end)
---
<details>
<summary>Dependabot commands and options</summary>
<br />
You can trigger Dependabot actions by commenting on this PR:
- `@dependabot rebase` will rebase this PR
- `@dependabot recreate` will recreate this PR, overwriting any edits
that have been made to it
- `@dependabot show <dependency name> ignore conditions` will show all
of the ignore conditions of the specified dependency
- `@dependabot ignore this major version` will close this PR and stop
Dependabot creating any more for this major version (unless you reopen
the PR or upgrade to it yourself)
- `@dependabot ignore this minor version` will close this PR and stop
Dependabot creating any more for this minor version (unless you reopen
the PR or upgrade to it yourself)
- `@dependabot ignore this dependency` will close this PR and stop
Dependabot creating any more for this dependency (unless you reopen the
PR or upgrade to it yourself)
</details>
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>1 parent 8cff165 commit 6ba36cf
3 files changed
Lines changed: 839 additions & 83 deletions
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
14 | 14 | | |
15 | 15 | | |
16 | 16 | | |
17 | | - | |
| 17 | + | |
18 | 18 | | |
19 | 19 | | |
20 | 20 | | |
| |||
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
14 | 14 | | |
15 | 15 | | |
16 | 16 | | |
17 | | - | |
| 17 | + | |
18 | 18 | | |
19 | 19 | | |
20 | 20 | | |
| |||
0 commit comments