feat(#2377): forge-webhook-triggered config-PR merge flow

Replace the request_merge_config_pr MCP tool with a Forgejo
pull_request webhook on the agent-configs org. Agents now open a
config PR normally; hive-c0re auto-queues the MergeConfigPr approval
from the webhook event — no extra tool call needed.

Changes:
- dashboard/webhook.rs: add POST /webhook/config-pr handler
  - parses Forgejo pull_request payload (opened/synchronize)
  - strips agent-configs/<agent> prefix to extract agent name
  - calls submit_merge_config_pr → queues dashboard approval card
  - always 200 to prevent Forgejo retries; errors logged at warn
- dashboard/mod.rs: wire /webhook/config-pr route
- forge/mod.rs: add ensure_config_pr_webhook() — idempotent org-level
  hook registration on agent-configs at startup; CONFIG_ORG now
  pub(crate) for webhook handler
- main.rs: call ensure_config_pr_webhook alongside knowledge webhook
- socket_server/config_approvals.rs: drop handle_request_merge_config_pr;
  make submit_merge_config_pr pub(crate) for webhook handler
- socket_server/mod.rs: re-export submit_merge_config_pr; drop dispatch arm
- hive-sh4re/src/lib.rs: remove AgentRequest::RequestMergeConfigPr
  wire type; drop from ToolGroup::Approvals tool list
- hive-ag3nt/src/mcp/: drop request_merge_config_pr tool + args struct
- docs: update approvals.md (webhook trigger), conventions.md (tool
  group), agent-hierarchy.md, tools/lifecycle.md

Hardening from #2375-merge-config-pr-hardening branch preserved:
- pr_is_open check at queue time (rejects closed/merged PRs)
- atomic fetched_sha INSERT via submit_kind(fetched_sha: Some(&sha))

Approve-handler machinery unchanged (run_merge_config_pr,
ff_push_to_main, fetch_pr_head_into_applied, mark_pr_merged).
This commit is contained in:
atlas 2026-07-11 10:50:24 +02:00 committed by mara
commit d5a81f9195
14 changed files with 255 additions and 163 deletions

View file

@ -99,7 +99,6 @@ Once enforcement lands the rules collapse into:
| `kill` / `start` / `restart` / `update` (any descendant) | any ancestor |
| `request_init_config` (spawn a new child) | any agent, child added under self |
| `request_apply_commit` (any descendant's config) | any ancestor |
| `request_merge_config_pr` (any descendant's forge config PR) | any ancestor |
| `get_logs` (any descendant) | any ancestor |
| moderate questions / reminders (cancel any open thread of a descendant) | any ancestor |
| `send` / `recv` routing | parent ↔ same-parent siblings ↔ self ↔ descendants; explicit allow-list for anyone else |

View file

@ -111,15 +111,20 @@ kind-specific payload carrier.
row (only `ApplyCommit` populates it). See the End-to-end flow
above.
- `MergeConfigPr` — the PR-based config flow's counterpart to
`ApplyCommit`. `commit_ref` stores the **PR number** (decimal),
and `fetched_sha` is the PR **head sha the operator reviewed**.
On approve, `run_merge_config_pr` re-reads the live PR head and
aborts if it drifted from `fetched_sha` (re-review), then fetches
that head into the applied repo, eval-verifies it, fast-forwards
the forge config repo's `main` to it (the merge), marks the PR
merged (best-effort — `main` is already there), and runs the same
shared deploy tail as `ApplyCommit` (`deploy_applied_target`).
Never a first spawn.
`ApplyCommit`. Triggered automatically: when an agent opens (or
force-pushes) a PR on its `agent-configs/<agent>` forge repo,
hive-c0re's `/webhook/config-pr` endpoint receives the Forgejo
pull_request event and queues this approval row. No MCP tool call
needed — the forge PR IS the request. `commit_ref` stores the
**PR number** (decimal), and `fetched_sha` is the PR **head sha
at queue time** (the "reviewed" sha). On approve,
`run_merge_config_pr` re-reads the live PR head and aborts if it
drifted from `fetched_sha` (submitter must push again to
re-trigger), then fetches that head into the applied repo,
eval-verifies it, fast-forwards the forge config repo's `main` to
it (the merge), marks the PR merged (best-effort — `main` is
already there), and runs the same shared deploy tail as
`ApplyCommit` (`deploy_applied_target`). Never a first spawn.
- `Spawn` — direct container creation under the default
`agent.nix` template. `commit_ref` is empty. Submitted via
`HostRequest::RequestSpawn` (operator-gated, the

View file

@ -325,7 +325,7 @@ binary flavor.
| `inbox` | `get_loose_ends`, `cancel_loose_end`, `remind`, `request_next_turn` |
| `execution` | vestigial — `mcp__bash__run` / `mcp__bash__status` are always available unconditionally via `extraMcpServers`; this group's entries expand to non-existent `mcp__hyperhive__run` / `mcp__hyperhive__status` and have no effect. See `docs/tools/bash.md`. |
| `lifecycle` | `kill`, `start`, `restart`, `update` *(privileged)* |
| `approvals` | `request_init_config`, `request_apply_commit`, `request_merge_config_pr`, `request_update_meta_inputs` *(privileged)* |
| `approvals` | `request_init_config`, `request_apply_commit`, `request_update_meta_inputs` *(privileged)* |
| `scheduling` | `request_schedule_prompt`, `fire_schedule_now`, `cancel_schedule`, `edit_schedule`, `list_schedules` *(privileged)* |
| `diagnostics` | `get_logs` *(privileged)* |

View file

@ -64,24 +64,6 @@ pinned commit.
rejected — the approval pins the exact commit). `agent` must be a
direct child. Topology-enforced.
### `request_merge_config_pr(agent, pr_number, description?)`
Submit an open PR on the agent's `agent-configs/<agent>` forge repo for
operator review and merge. The PR-based config flow's counterpart to
`request_apply_commit`: instead of pinning a commit sha from the proposed
repo, the submitter references an already-open forge PR.
hive-c0re fetches the PR head sha at submission time (the "reviewed" sha);
on operator approval it re-checks for drift, eval-verifies the commit,
fast-forwards the forge repo's `main` to the reviewed sha, marks the PR
merged, and rebuilds the agent container. If the PR head moves between
submission and approval the approve handler aborts — the submitter must
re-submit.
`agent` must be in the caller's subtree. The agent must already be fully
provisioned (applied repo present); this tool is not for first-spawn.
Requires the `approvals` tool group.
### `request_update_meta_inputs(inputs?, description?)`
Queue an approval to run `nix flake update [inputs...]` on the meta
@ -100,7 +82,6 @@ agents after the approval resolves.
| `list_containers` | No | All descendants |
| `request_init_config` | Yes (InitConfig) | New direct child only |
| `request_apply_commit` | Yes (ApplyCommit) | Direct children |
| `request_merge_config_pr` | Yes (MergeConfigPr) | Descendants |
| `request_update_meta_inputs` | Yes (MetaUpdate) | Meta flake (global) |
## See also

View file

@ -222,17 +222,6 @@ pub struct RequestApplyCommitArgs {
pub description: Option<String>,
}
#[derive(Debug, serde::Deserialize, schemars::JsonSchema)]
pub struct RequestMergeConfigPrArgs {
/// Logical agent name whose `agent-configs/<agent>` forge repo holds the PR.
pub agent: String,
/// Open PR index on the `agent-configs/<agent>` repo.
pub pr_number: u64,
/// Optional description shown on the dashboard approval card.
#[serde(default)]
pub description: Option<String>,
}
#[derive(Debug, serde::Deserialize, schemars::JsonSchema)]
pub struct UpdateMetaInputsArgs {
/// Flake input names to update (e.g. `["bitburner-agent", "nixpkgs"]`).

View file

@ -29,8 +29,8 @@ pub use args::{
AckUntilArgs, AgentGetLooseEndsArgs, AnswerArgs, AskArgs, CancelLooseEndArgs,
CancelScheduleArgs, CreateRepoArgs, EditScheduleArgs, FireScheduleNowArgs, GetAgentMetaArgs,
GetHostJournalArgs, GetLogsArgs, KillArgs, RecvArgs, RemindArgs, RequestApplyCommitArgs,
RequestInitConfigArgs, RequestMergeConfigPrArgs, RequestSchedulePromptArgs, RestartArgs,
SendArgs, SetStatusArgs, StartArgs, UpdateArgs, UpdateMetaInputsArgs,
RequestInitConfigArgs, RequestSchedulePromptArgs, RestartArgs, SendArgs, SetStatusArgs,
StartArgs, UpdateArgs, UpdateMetaInputsArgs,
};
pub use render::{
IDLE_WAIT_HINT, REDELIVERY_HINT, annotate_retries, format_ack, format_agent_meta, format_recv,
@ -757,47 +757,6 @@ impl AgentServer {
.await
}
// IMPORTANT: this tool is only available when the `approvals` tool group
// is configured for the agent. hive-c0re enforces both the tool-group check
// and topology: the target must be in the caller's subtree.
#[tool(
description = "Submit an open forge PR on `agent-configs/<agent>` for the operator \
to review and merge into the agent's running config. Requires the `approvals` \
tool group. `agent` must be in this agent's subtree. `pr_number` is the PR \
index on the `agent-configs/<agent>` repo. hive-c0re fetches the PR head sha \
at submission time (the drift-gate sha); if the PR head moves before the \
operator approves, the approve handler aborts without making any changes the \
submitter must re-submit. On approval hive-c0re eval-verifies the head, \
fast-forwards the forge repo's `main`, marks the PR merged, and rebuilds the \
agent container."
)]
async fn request_merge_config_pr(
&self,
Parameters(args): Parameters<RequestMergeConfigPrArgs>,
) -> String {
let log = format!("{args:?}");
let agent = args.agent.clone();
let pr_number = args.pr_number;
run_tool_envelope("request_merge_config_pr", log, async move {
let (resp, retries) = self
.dispatch(hive_sh4re::Request::RequestMergeConfigPr {
agent: args.agent,
pr_number: args.pr_number,
description: args.description,
})
.await;
annotate_retries(
format_ack(
resp,
"request_merge_config_pr",
format!("merge_config_pr approval queued for {agent} PR #{pr_number}"),
),
retries,
)
})
.await
}
// IMPORTANT: this tool is only available when the `lifecycle` tool group
// is granted to this agent. hive-c0re enforces the topology check
// server-side: the call is rejected unless `name` is a direct child.

View file

@ -160,9 +160,10 @@ pub async fn serve(port: u16, coord: Arc<Coordinator>) -> Result<()> {
post(schedules::post_rebuild_queue_cancel),
)
.route("/webhook/knowledge", post(webhook::post_webhook_knowledge))
.route("/webhook/config-pr", post(webhook::post_webhook_config_pr))
// Backend routes — the frontend calls these `/api/` paths. The
// transitional bare top-level aliases were removed once the
// frontend migrated. `/webhook/knowledge` keeps its own prefix
// frontend migrated. `/webhook/*` keeps its own prefix
// (forge-driven, not the SPA).
.route("/api/approve/{id}", post(approvals::post_approve))
.route("/api/deny/{id}", post(approvals::post_deny))

View file

@ -1,15 +1,27 @@
//! Forgejo push-webhook endpoint for the `internal/knowledge` repo.
//! Forgejo webhook endpoints.
//!
//! Loopback-only; on a push to `main` of the knowledge repo it triggers a
//! read-only `git pull` on the local clone so agents see up-to-date
//! documents on their next turn.
//! - **`/webhook/knowledge`** — push events on `internal/knowledge` trigger a
//! `git pull` on the local clone so agents see up-to-date docs.
//! - **`/webhook/config-pr`** — pull_request events on any `agent-configs/*`
//! repo queue a [`hive_sh4re::ApprovalKind::MergeConfigPr`] approval row
//! so the operator can review + approve the merge from the dashboard.
//!
//! Both endpoints are loopback-only (the axum listener binds
//! `127.0.0.1:<port>`) and have no signature verification (the risk is low:
//! loopback access implies host compromise already, and the config-PR path
//! still requires the operator to approve on the dashboard).
use axum::{
extract::State,
http::StatusCode,
response::{IntoResponse, Response},
};
use serde::Deserialize;
use super::AppState;
// ── knowledge webhook ──────────────────────────────────────────────────────────
/// Minimal Forgejo push-webhook payload — only the fields we care about.
#[derive(Deserialize)]
pub(super) struct PushWebhookPayload {
@ -62,3 +74,133 @@ pub(super) async fn post_webhook_knowledge(
});
(StatusCode::OK, "ok").into_response()
}
// ── config-PR webhook ──────────────────────────────────────────────────────────
/// Minimal Forgejo pull_request-webhook payload.
///
/// Forgejo fires this for actions: `opened`, `closed`, `reopened`,
/// `synchronize`, `assigned`, `unassigned`, `label_updated`,
/// `label_cleared`, `milestoned`, `demilestoned`, `review_requested`,
/// `review_request_removed`, `auto_merge_enabled`, `auto_merge_disabled`.
/// We only act on `opened` and `synchronize`.
#[derive(Deserialize)]
pub(super) struct PrWebhookPayload {
/// What triggered this event (`opened`, `closed`, `synchronize`, …).
action: Option<String>,
/// PR index on the repo.
number: Option<u64>,
pull_request: Option<PrWebhookPr>,
repository: Option<PrWebhookRepo>,
}
#[derive(Deserialize)]
struct PrWebhookPr {
head: Option<PrWebhookHead>,
}
#[derive(Deserialize)]
struct PrWebhookHead {
sha: Option<String>,
}
#[derive(Deserialize)]
struct PrWebhookRepo {
full_name: Option<String>,
}
/// POST `/webhook/config-pr` — Forgejo pull_request webhook for
/// `agent-configs/*` repos.
///
/// On `opened` or `synchronize` for an `agent-configs/<agent>` PR:
/// fetches the current PR head sha, queues a `MergeConfigPr` approval row,
/// and emits the `ApprovalAdded` event so the dashboard card appears
/// immediately.
///
/// All other actions (closed, label changes, etc.) are silently ignored —
/// the operator can deny a pending approval if the PR is later closed.
///
/// Always returns HTTP 200 (even on queue failure) so Forgejo does not
/// retry the delivery. Failures are logged at `warn` level.
///
/// Expected Forgejo webhook configuration:
/// - URL: `http://127.0.0.1:<dashboard_port>/webhook/config-pr`
/// - Content type: `application/json`
/// - Events: "Pull Request" only
/// - Organisation: `agent-configs` (org-level hook covers all config repos)
///
/// hive-c0re registers this hook automatically at startup via
/// [`crate::forge::ensure_config_pr_webhook`].
pub(super) async fn post_webhook_config_pr(
State(state): State<AppState>,
axum::extract::Json(payload): axum::extract::Json<PrWebhookPayload>,
) -> Response {
let action = payload.action.as_deref().unwrap_or("");
// Only act on newly-opened or force-updated PRs.
if action != "opened" && action != "synchronize" {
tracing::debug!(action, "webhook/config-pr: ignoring action");
return (StatusCode::OK, "ignored").into_response();
}
let full_name = payload
.repository
.as_ref()
.and_then(|r| r.full_name.as_deref())
.unwrap_or("");
// Expect `agent-configs/<agent>`.
let agent = match full_name.strip_prefix(&format!("{}/", crate::forge::CONFIG_ORG)) {
Some(name) if !name.is_empty() && !name.contains('/') => name,
_ => {
tracing::debug!(
full_name,
"webhook/config-pr: ignoring non-config-repo event"
);
return (StatusCode::OK, "ignored").into_response();
}
};
let pr_number = match payload.number {
Some(n) if n > 0 => n,
_ => {
tracing::warn!(full_name, "webhook/config-pr: missing or zero PR number");
return (StatusCode::OK, "ignored").into_response();
}
};
// The payload already carries the head sha — use it as an early hint for
// logging, but the canonical sha comes from `submit_merge_config_pr`'s
// fresh forge API call so we don't trust a potentially-stale payload sha.
let payload_sha = payload
.pull_request
.as_ref()
.and_then(|pr| pr.head.as_ref())
.and_then(|h| h.sha.as_deref())
.unwrap_or("<unknown>");
tracing::info!(
%full_name, %agent, %pr_number, %payload_sha, %action,
"webhook/config-pr: queuing MergeConfigPr approval"
);
// Queue the approval. The description surfaces the action and PR number
// on the dashboard card so the operator has context without opening the
// forge PR.
let description = format!("PR #{pr_number} on {full_name} ({action})");
if let Err(e) = crate::socket_server::submit_merge_config_pr(
&state.coord,
agent,
pr_number,
Some(&description),
"forge", // submitter — identifies the webhook path in the audit trail
)
.await
{
tracing::warn!(
%agent, %pr_number, error = ?e,
"webhook/config-pr: failed to queue MergeConfigPr approval"
);
}
(StatusCode::OK, "ok").into_response()
}

View file

@ -69,7 +69,7 @@ pub(crate) fn forge_git_url(token: &str, repo: &str) -> String {
/// reach *another* agent's config. `main` is fast-forward-only — hive-c0re
/// never force-pushes; the `push_config` mirror runs best-effort until the
/// PR-merge flow retires it.
const CONFIG_ORG: &str = "agent-configs";
pub(crate) const CONFIG_ORG: &str = "agent-configs";
/// Forgejo org hosting the operator-curated shared docs/skills repo
/// that every agent gets read-only access to. Agents use it as a
/// common reference without the operator having to bake content into
@ -294,3 +294,70 @@ pub async fn ensure_all() {
sync_agent(name, core_token.as_deref()).await;
}
}
/// Ensure a Forgejo pull_request org-webhook for `agent-configs` exists and
/// points at hive-c0re's `/webhook/config-pr` endpoint. Idempotent — lists
/// existing hooks first and skips creation when one is already targeting the
/// correct URL. `dashboard_port` is the TCP port hive-c0re's dashboard listens
/// on (default 7000); the webhook URL is
/// `http://127.0.0.1:<port>/webhook/config-pr`.
///
/// An org-level hook covers every repo in `agent-configs` automatically,
/// so no per-repo setup is needed as new agents are provisioned.
///
/// Called at startup alongside `knowledge::ensure_webhook`. No-op when the
/// core token is absent (forge not yet provisioned).
pub async fn ensure_config_pr_webhook(core_token: &str, dashboard_port: u16) -> Result<()> {
use forgejo_api::structs::{CreateHookOption, CreateHookOptionConfig, CreateHookOptionType};
use std::collections::BTreeMap;
const HTTP_TIMEOUT: std::time::Duration = std::time::Duration::from_secs(10);
let target_url = format!("http://127.0.0.1:{dashboard_port}/webhook/config-pr");
let client = api(core_token)?;
// List existing org hooks — skip creation if ours is already there.
// Best-effort: a listing failure falls through to the create attempt.
let listed = tokio::time::timeout(HTTP_TIMEOUT, client.org_list_hooks(CONFIG_ORG).all())
.await
.map_err(anyhow::Error::from)
.and_then(|r| r.map_err(anyhow::Error::from));
match listed {
Ok(hooks) => {
let already_exists = hooks.iter().any(|h| {
h.config
.as_ref()
.and_then(|c| c.get("url"))
.map(String::as_str)
== Some(target_url.as_str())
});
if already_exists {
tracing::debug!(%target_url, "forge: config-pr webhook already configured");
return Ok(());
}
}
Err(e) => {
tracing::debug!(error = %e, "forge: listing config-pr hooks failed; attempting create");
}
}
let hook = CreateHookOption {
active: Some(true),
authorization_header: None,
branch_filter: None,
config: CreateHookOptionConfig {
content_type: "json".to_owned(),
url: Url::parse(&target_url).context("parse config-pr webhook target url")?,
additional: BTreeMap::new(),
},
events: Some(vec!["pull_request".to_owned()]),
r#type: CreateHookOptionType::Forgejo,
};
tokio::time::timeout(HTTP_TIMEOUT, client.org_create_hook(CONFIG_ORG, hook))
.await
.map_err(anyhow::Error::from)
.and_then(|r| r.map_err(anyhow::Error::from))
.with_context(|| format!("create config-pr webhook on org {CONFIG_ORG}"))?;
tracing::info!(%target_url, "forge: config-pr webhook created on org {CONFIG_ORG}");
Ok(())
}

View file

@ -303,17 +303,22 @@ async fn cmd_serve(
tokio::spawn(async move {
forge::ensure_all().await;
});
// Knowledge webhook setup: ensure the Forgejo push webhook for
// `internal/knowledge` exists so `pull()` fires on merge. Runs
// after forge::ensure_all so the core token + repo are present.
// Webhook setup: ensure Forgejo webhooks are registered for both
// `internal/knowledge` (push → git pull) and the `agent-configs` org
// (pull_request → queue MergeConfigPr approval). Both run after
// forge::ensure_all so the core token + repos + org are present.
// No-op when the core token or forge are absent.
let webhook_port = dashboard_port;
tokio::spawn(async move {
if let Some(token) = forge::core_token()
&& let Err(e) = knowledge::ensure_webhook(&token, webhook_port).await
{
let Some(token) = forge::core_token() else {
return;
};
if let Err(e) = knowledge::ensure_webhook(&token, webhook_port).await {
tracing::warn!(error = ?e, "knowledge: ensure_webhook failed");
}
if let Err(e) = forge::ensure_config_pr_webhook(&token, webhook_port).await {
tracing::warn!(error = ?e, "forge: ensure_config_pr_webhook failed");
}
});
// Knowledge periodic pull: hourly fallback in case the webhook is
// missed (e.g. hive-c0re was down during a push). First fires at

View file

@ -1,8 +1,13 @@
//! Config-approval request handlers: `RequestInitConfig` /
//! `RequestApplyCommit` / `RequestMergeConfigPr` / `RequestUpdateMetaInputs`,
//! `RequestApplyCommit` / `RequestUpdateMetaInputs`,
//! plus the shared submit helpers (`submit_init_config` / `submit_apply_commit`
//! / `submit_merge_config_pr`) and the commit-sha shape check
//! (`validate_commit_ref`).
//!
//! `submit_merge_config_pr` is called from the dashboard webhook handler
//! (`dashboard::webhook`) — agents no longer need an MCP tool for this;
//! opening a config PR on `agent-configs/<agent>` is enough to trigger
//! hive-c0re's webhook-driven queue path.
use std::sync::Arc;
@ -107,35 +112,6 @@ pub(super) fn handle_request_update_meta_inputs(
AgentResponse::Ok
}
/// `RequestMergeConfigPr` — queue a `MergeConfigPr` approval for a PR on an
/// agent's `agent-configs/<agent>` forge repo. The target must be in the
/// caller's subtree. hive-c0re fetches the PR head sha at submission time;
/// that sha is stored as `fetched_sha` and forms the drift gate in the
/// approve handler: if the PR head moves between submission and approval,
/// the approve handler aborts without making any changes.
pub(super) async fn handle_request_merge_config_pr(
coord: &Arc<Coordinator>,
agent: &str,
target_agent: &str,
pr_number: u64,
description: Option<&str>,
) -> AgentResponse {
if let Some(err) = super::require_descendant(agent, target_agent, "request_merge_config_pr for")
{
return err;
}
tracing::info!(%agent, %target_agent, %pr_number, "request_merge_config_pr");
match submit_merge_config_pr(coord, target_agent, pr_number, description, agent).await {
Ok(id) => {
tracing::info!(%id, %target_agent, %pr_number, "merge_config_pr approval queued");
AgentResponse::Ok
}
Err(e) => AgentResponse::Err {
message: format!("{e:#}"),
},
}
}
/// Submit-time half of the PR-merge flow: fetch the PR head sha from the
/// forge, queue the approval row, and emit the `approval_added` event so the
/// dashboard shows the pending card immediately.
@ -146,7 +122,7 @@ pub(super) async fn handle_request_merge_config_pr(
/// this does NOT fetch the commit into the applied repo at submission time
/// (that happens inside the approve handler, step 2, after the drift check).
/// No flake pre-flight either — eval-verify happens at approval time too.
async fn submit_merge_config_pr(
pub(crate) async fn submit_merge_config_pr(
coord: &Arc<Coordinator>,
agent: &str,
pr_number: u64,
@ -172,7 +148,7 @@ async fn submit_merge_config_pr(
{
anyhow::bail!(
"PR #{pr_number} on {repo} is closed or already merged — \
request_merge_config_pr requires an open PR"
merge_config_pr requires an open PR"
);
}
// Fetch the current PR head sha — becomes the "reviewed" sha.

View file

@ -25,12 +25,12 @@ mod lifecycle_handlers;
mod reminders;
mod schedules;
pub(crate) use config_approvals::{submit_init_config, submit_merge_config_pr};
pub(crate) use schedules::filter_ghost_schedule_targets;
pub use schedules::schedule_to_wire_public;
use config_approvals::{
handle_request_apply_commit, handle_request_init_config, handle_request_merge_config_pr,
handle_request_update_meta_inputs,
handle_request_apply_commit, handle_request_init_config, handle_request_update_meta_inputs,
};
use lifecycle_handlers::{
handle_kill, handle_list_descendants, handle_restart, handle_start, handle_update,
@ -579,23 +579,6 @@ async fn dispatch(req: &AgentRequest, agent: &str, coord: &Arc<Coordinator>) ->
)
.await
}
AgentRequest::RequestMergeConfigPr {
agent: target_agent,
pr_number,
description,
} => {
if let Some(err) = require_group(agent, "approvals", "request merge_config_pr") {
return err;
}
handle_request_merge_config_pr(
coord,
agent,
target_agent,
*pr_number,
description.as_deref(),
)
.await
}
// Agent-state queries: own subtree is free; other agents + the
// hive-wide `"*"` sweep require `QueryAgentState`.
AgentRequest::GetLooseEnds { agent: target } => {

View file

@ -788,20 +788,6 @@ pub enum Request {
#[serde(default, skip_serializing_if = "Option::is_none")]
description: Option<String>,
},
/// *(privileged)* Submit a forge config PR for the operator to review and
/// merge. `agent` is the child whose `agent-configs/<agent>` forge repo
/// holds the PR; `pr_number` is the open PR index on that repo.
/// hive-c0re fetches the PR head sha at submission time (the "reviewed"
/// sha for the drift gate) and queues a `MergeConfigPr` approval. On
/// approval, hive-c0re re-verifies the head hasn't drifted, eval-verifies
/// the commit, fast-forwards the forge repo's `main`, marks the PR merged,
/// and rebuilds the agent container.
RequestMergeConfigPr {
agent: String,
pr_number: u64,
#[serde(default, skip_serializing_if = "Option::is_none")]
description: Option<String>,
},
/// *(privileged)* Fetch recent journal lines for a sub-agent container.
GetLogs {
agent: String,
@ -1215,7 +1201,6 @@ impl ToolGroup {
Self::Approvals => &[
"request_init_config",
"request_apply_commit",
"request_merge_config_pr",
"request_update_meta_inputs",
],
Self::Scheduling => &[
@ -1318,7 +1303,7 @@ impl ToolGroup {
"kill, start, restart, update, list_containers — container lifecycle (privileged)"
}
Self::Approvals => {
"request_init_config, request_apply_commit, request_merge_config_pr, request_update_meta_inputs — config change flow (privileged)"
"request_init_config, request_apply_commit, request_update_meta_inputs — config change flow (privileged)"
}
Self::Scheduling => {
"request_schedule_prompt and related — operator-visible scheduled prompts (privileged)"

BIN
libnull.rlib Normal file

Binary file not shown.