feat(coding-agents): first-class pi support, sharing one extension adapter and installer with Prime Agent

README.md

@@ -254,7 +254,7 @@ npx @vectorize-io/hindsight-coding-agents install all          # every detected
 npx @vectorize-io/hindsight-coding-agents install claude-code  # or just one
 ```
 
-Supports Claude Code, Codex CLI, Cursor CLI, GitHub Copilot CLI, opencode, Kilo CLI, Cline CLI, Antigravity CLI, Devin CLI, Prime Agent, Grok Build and DeepSeek Harness. Ingestion is automatic — there is no setup command. See the [coding agents integration](https://hindsight.vectorize.io/sdks/integrations/coding-agents).
+Supports Claude Code, Codex CLI, Cursor CLI, GitHub Copilot CLI, opencode, Kilo CLI, Cline CLI, Antigravity CLI, Devin CLI, pi, Prime Agent, Grok Build and DeepSeek Harness. Ingestion is automatic — there is no setup command. See the [coding agents integration](https://hindsight.vectorize.io/sdks/integrations/coding-agents).
 
 ### MCP Server
 

hindsight-control-plane/src/lib/harness-logo.ts

@@ -13,7 +13,9 @@
  * The ids below are exactly the ones that integration emits — see its
  * `src/harness/hook-lifecycle.ts` (one HookSpec per hook-driven agent) and
  * `src/harness/registry.ts` (the persistent-plugin agents, whose id is the
- * `createPluginEntry(...)` argument of their entrypoint). Do not add entries for
+ * argument their entrypoint passes to the shared factory it builds on —
+ * `createPluginEntry(...)` for the opencode family, `createPiExtension(...)` for
+ * the pi family). Do not add entries for
  * agents that cannot appear yet: an id nothing writes is a logo nothing renders.
  * When that integration gains a harness, add it here in the same change — drop
  * its icon in `public/img/harness/` (copied from
@@ -87,6 +89,10 @@ export const HARNESS_LOGO_REGISTRY: Record<string, HarnessLogoEntry> = {
   "grok-build": { id: "grok-build", label: "Grok Build", src: "/img/harness/grok-build.svg" },
   kilo: { id: "kilo", label: "Kilo CLI", src: "/img/harness/kilo.svg" },
   opencode: { id: "opencode", label: "OpenCode", src: "/img/harness/opencode.png" },
+  // pi ships no brand mark of its own, so this is the letter the agent is named
+  // for, drawn in the same monochrome style as the Prime Agent mark below (the
+  // fork of pi) rather than a logo invented for it.
+  pi: { id: "pi", label: "pi", src: "/img/harness/pi.svg", invertOnDark: true },
   "prime-agent": {
     id: "prime-agent",
     label: "Prime Agent",

hindsight-control-plane/tests/lib/harness-logo.test.ts

@@ -36,8 +36,9 @@ describe("documentHarness", () => {
 describe("resolveHarnessLogo", () => {
   // The exact set hindsight-coding-agents emits: one per HookSpec in
   // src/harness/hook-lifecycle.ts, plus the persistent-plugin harnesses whose id
-  // is their entrypoint's createPluginEntry(...) argument (opencode, kilo,
-  // cline-cli). The registry must cover these and nothing speculative.
+  // is their entrypoint's argument to the shared factory they use (opencode,
+  // kilo, cline-cli, pi, prime-agent). The registry must cover these and nothing
+  // speculative.
   const EMITTED_HARNESSES = [
     "antigravity-cli",
     "claude-code",
@@ -50,6 +51,7 @@ describe("resolveHarnessLogo", () => {
     "grok-build",
     "kilo",
     "opencode",
+    "pi",
     "prime-agent",
   ];
   // Ids the integration used to emit. Kept so documents already retained under

hindsight-docs/docs-integrations/coding-agents.md

@@ -9,7 +9,7 @@ description: "One Hindsight memory plugin for coding agents — per-repo memory
 
 Long-term project memory for **coding agents**, backed by [Hindsight](https://vectorize.io/hindsight).
 One package, several agents: a shared reflect-and-inject core with a thin entry point per agent
-(**opencode**, **Kilo CLI**, **Cline CLI**, **Prime Agent**, **DeepSeek Harness**, **Claude Code**, **Codex CLI**, **Antigravity CLI**, **Cursor CLI**, **GitHub Copilot CLI**, **Devin CLI**, **Grok Build**). Ingestion is fully
+(**opencode**, **Kilo CLI**, **Cline CLI**, **pi**, **Prime Agent**, **DeepSeek Harness**, **Claude Code**, **Codex CLI**, **Antigravity CLI**, **Cursor CLI**, **GitHub Copilot CLI**, **Devin CLI**, **Grok Build**). Ingestion is fully
 automatic — there is no setup command: a repo's git history and conversations flow into its memory
 bank in the background as you work.
 
@@ -120,14 +120,32 @@ npx @vectorize-io/hindsight-coding-agents install cline-cli
 
 A native plugin via `cline plugin install`, plus MCP and the companion skill.
 
+#### <img src="/img/harness/pi.svg" alt="" width="20" height="20" /> pi
+
+```bash
+npx @vectorize-io/hindsight-coding-agents install pi
+```
+
+An extension entry in `~/.pi/agent/settings.json`, plus the companion skill in
+`~/.pi/agent/skills` — native tools, no MCP needed. Use this command
+rather than `pi install npm:@vectorize-io/hindsight-coding-agents`: pi and Prime Agent read the same
+`pi` key of a package's `package.json`, which can only name one entry, and it names Prime Agent's —
+so on the package route pi would report itself as `prime-agent`, taking that harness's config
+section and stamping its documents with the wrong agent.
+
 #### <img src="/img/harness/prime-agent.svg" alt="" width="20" height="20" /> Prime Agent
 
 ```bash
 npx @vectorize-io/hindsight-coding-agents install prime-agent
 ```
 
-An extension entry in `~/.prime/agent/settings.json` — native tools, no MCP needed — plus the
-companion skill in `~/.prime/agent/skills`.
+Prime Agent is a fork of pi, so it is wired the same way: an extension entry, here in
+`~/.prime/agent/settings.json`, plus the companion skill in `~/.prime/agent/skills` — native tools,
+no MCP needed. Installing both is fine and expected:
+each host loads its own entry from its own settings file, and like every other pair of agents they
+**share one bank per repo** (the default `coding-agent::{gitProject}`), so what you tell pi is there
+when you open Prime Agent. Separate entries are what keeps each side attributable — its own
+`harnesses.<name>` config section, and its own agent stamped on every document it retains.
 
 #### <img src="/img/harness/dsh.svg" alt="" width="20" height="20" /> DeepSeek Harness
 
@@ -212,7 +230,8 @@ transcripts are needed either way; going through them directly is simply the sho
 **How sessions are matched.** A conversation is imported only when the session itself records the
 directory it ran in — never inferred from a file or folder name. Claude Code writes that directory
 on its entries, Codex in its `session_meta` header and DeepSeek Harness in its session-log header,
-so all three can be attributed exactly, including sessions started in a subdirectory of the repo. Guessing was tempting (Claude names its history
+and pi and Prime Agent in their session header, so all five can be attributed exactly, including
+sessions started in a subdirectory of the repo. Guessing was tempting (Claude names its history
 folders after the project path) but unsafe: `/` and `.` both encode to `-`, so `repo-sub` is either
 the subdirectory `repo/sub` or an unrelated sibling repo — and a wrong guess files someone else's
 conversation into your bank. Sessions that record nothing are skipped and the count is reported.
@@ -320,7 +339,7 @@ what reads it:
 | host                                                                                                        | reads the file                                                      | an edit applies            |
 | ----------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------- | -------------------------- |
 | hook harnesses (Claude Code, Codex CLI, Cursor CLI, GitHub Copilot CLI, Grok Build, Antigravity CLI, Devin) | once per hook invocation — each hook is its own short-lived process | on your next prompt        |
-| persistent plugins (opencode, Kilo CLI, Cline CLI, Prime Agent, DeepSeek Harness)                           | once per workspace, when the host loads the plugin                  | after restarting the agent |
+| persistent plugins (opencode, Kilo CLI, Cline CLI, pi, Prime Agent, DeepSeek Harness)                       | once per workspace, when the host loads the plugin                  | after restarting the agent |
 | the MCP server behind the `hindsight_*` tools                                                               | once at startup                                                     | in your next session       |
 
 `apiToken` is the exception. Every host re-reads it when the server rejects a request, so enabling

hindsight-integrations/coding-agents/README.md

@@ -2,7 +2,7 @@
 
 Long-term project memory for **coding agents**, backed by [Hindsight](https://vectorize.io/hindsight).
 One package, several agents: a shared reflect-and-inject core with a thin entry point per agent
-(**opencode**, **Kilo CLI**, **Cline CLI**, **Prime Agent**, **DeepSeek Harness**, **Claude Code**, **Codex CLI**, **Antigravity CLI**, **Cursor CLI**, **GitHub Copilot CLI**, **Devin CLI**, **Grok Build**). Ingestion is fully
+(**opencode**, **Kilo CLI**, **Cline CLI**, **pi**, **Prime Agent**, **DeepSeek Harness**, **Claude Code**, **Codex CLI**, **Antigravity CLI**, **Cursor CLI**, **GitHub Copilot CLI**, **Devin CLI**, **Grok Build**). Ingestion is fully
 automatic — there is no setup command: a repo's git history and conversations flow into its memory
 bank in the background as you work.
 
@@ -117,14 +117,32 @@ npx @vectorize-io/hindsight-coding-agents install cline-cli
 
 A native plugin via `cline plugin install`, plus MCP and the companion skill.
 
+#### <img src="https://hindsight.vectorize.io/img/harness/pi.svg" alt="" width="20" height="20" /> pi
+
+```bash
+npx @vectorize-io/hindsight-coding-agents install pi
+```
+
+An extension entry in `~/.pi/agent/settings.json`, plus the companion skill in
+`~/.pi/agent/skills` — native tools, no MCP needed. Use this command
+rather than `pi install npm:@vectorize-io/hindsight-coding-agents`: pi and Prime Agent read the same
+`pi` key of a package's `package.json`, which can only name one entry, and it names Prime Agent's —
+so on the package route pi would report itself as `prime-agent`, taking that harness's config
+section and stamping its documents with the wrong agent.
+
 #### <img src="https://hindsight.vectorize.io/img/harness/prime-agent.svg" alt="" width="20" height="20" /> Prime Agent
 
 ```bash
 npx @vectorize-io/hindsight-coding-agents install prime-agent
 ```
 
-An extension entry in `~/.prime/agent/settings.json` — native tools, no MCP needed — plus the
-companion skill in `~/.prime/agent/skills`.
+Prime Agent is a fork of pi, so it is wired the same way: an extension entry, here in
+`~/.prime/agent/settings.json`, plus the companion skill in `~/.prime/agent/skills` — native tools,
+no MCP needed. Installing both is fine and expected:
+each host loads its own entry from its own settings file, and like every other pair of agents they
+**share one bank per repo** (the default `coding-agent::{gitProject}`), so what you tell pi is there
+when you open Prime Agent. Separate entries are what keeps each side attributable — its own
+`harnesses.<name>` config section, and its own agent stamped on every document it retains.
 
 #### <img src="https://hindsight.vectorize.io/img/harness/dsh.svg" alt="" width="20" height="20" /> DeepSeek Harness
 
@@ -209,7 +227,8 @@ transcripts are needed either way; going through them directly is simply the sho
 **How sessions are matched.** A conversation is imported only when the session itself records the
 directory it ran in — never inferred from a file or folder name. Claude Code writes that directory
 on its entries, Codex in its `session_meta` header and DeepSeek Harness in its session-log header,
-so all three can be attributed exactly, including sessions started in a subdirectory of the repo. Guessing was tempting (Claude names its history
+and pi and Prime Agent in their session header, so all five can be attributed exactly, including
+sessions started in a subdirectory of the repo. Guessing was tempting (Claude names its history
 folders after the project path) but unsafe: `/` and `.` both encode to `-`, so `repo-sub` is either
 the subdirectory `repo/sub` or an unrelated sibling repo — and a wrong guess files someone else's
 conversation into your bank. Sessions that record nothing are skipped and the count is reported.
@@ -323,7 +342,7 @@ what reads it:
 | host                                                                                                        | reads the file                                                      | an edit applies            |
 | ----------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------- | -------------------------- |
 | hook harnesses (Claude Code, Codex CLI, Cursor CLI, GitHub Copilot CLI, Grok Build, Antigravity CLI, Devin) | once per hook invocation — each hook is its own short-lived process | on your next prompt        |
-| persistent plugins (opencode, Kilo CLI, Cline CLI, Prime Agent, DeepSeek Harness)                           | once per workspace, when the host loads the plugin                  | after restarting the agent |
+| persistent plugins (opencode, Kilo CLI, Cline CLI, pi, Prime Agent, DeepSeek Harness)                       | once per workspace, when the host loads the plugin                  | after restarting the agent |
 | the MCP server behind the `hindsight_*` tools                                                               | once at startup                                                     | in your next session       |
 
 `apiToken` is the exception. Every host re-reads it when the server rejects a request, so enabling

hindsight-integrations/coding-agents/e2e/Dockerfile.pi

@@ -0,0 +1,3 @@
+FROM hindsight-coding-agents-e2e-base
+ARG VERSION=latest
+RUN npm install --global "@earendil-works/pi-coding-agent@${VERSION}" && command -v pi

hindsight-integrations/coding-agents/skill/SKILL.md

@@ -125,7 +125,7 @@ what reads it:
 | host                                                                                                        | reads the file                                                      | an edit applies            |
 | ----------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------- | -------------------------- |
 | hook harnesses (Claude Code, Codex CLI, Cursor CLI, GitHub Copilot CLI, Grok Build, Antigravity CLI, Devin) | once per hook invocation — each hook is its own short-lived process | on your next prompt        |
-| persistent plugins (opencode, Kilo CLI, Cline CLI, Prime Agent, DeepSeek Harness)                           | once per workspace, when the host loads the plugin                  | after restarting the agent |
+| persistent plugins (opencode, Kilo CLI, Cline CLI, pi, Prime Agent, DeepSeek Harness)                       | once per workspace, when the host loads the plugin                  | after restarting the agent |
 | the MCP server behind the `hindsight_*` tools                                                               | once at startup                                                     | in your next session       |
 
 `apiToken` is the exception. Every host re-reads it when the server rejects a request, so enabling