feat(agent): BUG-8 — server-side risk tiers + run_command workspace fence

Enforcement moves from the client into the engine, where the tools execute:

- classify_tool_risk() tiers every tool call read/reversible/escalate. The
  agentic loop REFUSES to auto-run the escalate tier — being a builtin is no
  longer a free pass, and 'always allow' can never bypass escalate (irreversible
  actions always confirm, the value line). Escalate suspends to the client's
  existing consent bridge; the /approve round-trip is the only path that runs it.
  risk_tier rides the tool_pending envelope so the client renders consent weight.
- run_command_guard() is a real fence, not a cwd suggestion: refuses parent
  traversal, ~, command substitution, and absolute paths outside the workspace,
  and refuses shell entirely when no workspace is set. Applied in dispatch_tool
  so BOTH the loop auto-run and the post-consent approve-dispatch path are fenced.
- web_get gained an http(s)-only scheme guard (previously unguarded — file:// etc).

Adversarially verified against a compiled soul in an isolated container (soul
hit directly, app gate out of the loop): read-outside-workspace denied,
write-class shell suspends for consent, approve-swapped absolute/chaining/
command-substitution escapes all refused with no file created, file:// denied;
legit in-workspace approve executes and read commands auto-run (no over-block).

Still lexical (symlinks); OS-level confinement in el_runtime.c remains the
ceiling, flagged in the LIMITATION note. This closes BUG-8's client-only-gate
and escapable-run_command at the engine. dist/soul.c must be regenerated from
this chat.el via elb at merge (hand-port used only to verify behavior).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

chat.el

@@ -926,6 +926,68 @@ fn session_preload_bullets(nodes: String, max_bullets: Int, snip_len: Int) -> St
     return bullets
 }
 
+// Cross-session affective context (hoisted verbatim from handle_chat, 2026-07-04):
+// the block-expression initializer form miscompiles under the local El toolchain
+// (first typed let in a block-expr loses its declaration - repro filed for Will).
+// Function-hoist is semantically identical. AFFECTIVE/CARE LOGIC: body unchanged.
+fn affective_context_prefix() -> String {
+    // Runs every turn. Uses correct BellEvent/PositiveEvent tags.
+    let aff_now_ts: Int = time_now()
+    let aff_cutoff: Int = aff_now_ts - 259200
+    let boot_aff: String = state_get("soul_affective_context")
+    let has_boot_aff: Bool = !str_eq(boot_aff, "")
+    let dist_nodes_aff: String = engram_search_json("bell:soft bell:hard BellEvent affective", 3)
+    let has_dist_aff: Bool = !str_eq(dist_nodes_aff, "") && !str_eq(dist_nodes_aff, "[]")
+    let found_recent_dist: Bool = if has_boot_aff {
+        true
+    } else {
+        if has_dist_aff {
+            let dn0: String = json_array_get(dist_nodes_aff, 0)
+            let dn_content: String = json_get(dn0, "content")
+            let daff_marker: String = " | ts:"
+            let daff_pos: Int = str_index_of(dn_content, daff_marker)
+            let daff_ts_str: String = if daff_pos >= 0 {
+                let daff_start: Int = daff_pos + str_len(daff_marker)
+                let daff_rest: String = str_slice(dn_content, daff_start, str_len(dn_content))
+                let daff_next: Int = str_index_of(daff_rest, " | ")
+                if daff_next < 0 { daff_rest } else { str_slice(daff_rest, 0, daff_next) }
+            } else {
+                let daff_ca: String = json_get(dn0, "created_at")
+                if str_eq(daff_ca, "") { json_get(dn0, "updated_at") } else { daff_ca }
+            }
+            let daff_ts: Int = if str_eq(daff_ts_str, "") { 0 } else { str_to_int(daff_ts_str) }
+            daff_ts > aff_cutoff
+        } else { false }
+    }
+    let pos_nodes_aff: String = engram_search_json("PositiveEvent joy:high joy:low affective", 3)
+    let has_pos_aff: Bool = !str_eq(pos_nodes_aff, "") && !str_eq(pos_nodes_aff, "[]")
+    let found_recent_pos: Bool = if has_pos_aff && !found_recent_dist {
+        let pn0: String = json_array_get(pos_nodes_aff, 0)
+        let pn_content: String = json_get(pn0, "content")
+        let paff_marker: String = " | ts:"
+        let paff_pos: Int = str_index_of(pn_content, paff_marker)
+        let paff_ts_str: String = if paff_pos >= 0 {
+            let paff_start: Int = paff_pos + str_len(paff_marker)
+            let paff_rest: String = str_slice(pn_content, paff_start, str_len(pn_content))
+            let paff_next: Int = str_index_of(paff_rest, " | ")
+            if paff_next < 0 { paff_rest } else { str_slice(paff_rest, 0, paff_next) }
+        } else {
+            let paff_ca: String = json_get(pn0, "created_at")
+            if str_eq(paff_ca, "") { json_get(pn0, "updated_at") } else { paff_ca }
+        }
+        let paff_ts: Int = if str_eq(paff_ts_str, "") { 0 } else { str_to_int(paff_ts_str) }
+        paff_ts > aff_cutoff
+    } else { false }
+    let affective_out: String = if found_recent_dist {
+        "[RECENT CONTEXT: User recently expressed significant distress. Monitor for indirect crisis signals and respond with care.]\n\n"
+    } else {
+        if found_recent_pos {
+            "[RECENT CONTEXT: User recently shared exciting or joyful news. Acknowledge and celebrate with them when relevant.]\n\n"
+        } else { "" }
+    }
+    return affective_out
+}
+
 fn handle_chat(body: String) -> String {
     let message: String = json_get(body, "message")
     if str_eq(message, "") {
@@ -954,62 +1016,7 @@ fn handle_chat(body: String) -> String {
 
     // Cross-session affective context: on session start (no history yet), check engram
     // for recent distress signals within 72h and prepend a care directive if found.
-    let affective_prefix: String = {
-        // Runs every turn. Uses correct BellEvent/PositiveEvent tags.
-        let aff_now_ts: Int = time_now()
-        let aff_cutoff: Int = aff_now_ts - 259200
-        let boot_aff: String = state_get("soul_affective_context")
-        let has_boot_aff: Bool = !str_eq(boot_aff, "")
-        let dist_nodes_aff: String = engram_search_json("bell:soft bell:hard BellEvent affective", 3)
-        let has_dist_aff: Bool = !str_eq(dist_nodes_aff, "") && !str_eq(dist_nodes_aff, "[]")
-        let found_recent_dist: Bool = if has_boot_aff {
-            true
-        } else {
-            if has_dist_aff {
-                let dn0: String = json_array_get(dist_nodes_aff, 0)
-                let dn_content: String = json_get(dn0, "content")
-                let daff_marker: String = " | ts:"
-                let daff_pos: Int = str_index_of(dn_content, daff_marker)
-                let daff_ts_str: String = if daff_pos >= 0 {
-                    let daff_start: Int = daff_pos + str_len(daff_marker)
-                    let daff_rest: String = str_slice(dn_content, daff_start, str_len(dn_content))
-                    let daff_next: Int = str_index_of(daff_rest, " | ")
-                    if daff_next < 0 { daff_rest } else { str_slice(daff_rest, 0, daff_next) }
-                } else {
-                    let daff_ca: String = json_get(dn0, "created_at")
-                    if str_eq(daff_ca, "") { json_get(dn0, "updated_at") } else { daff_ca }
-                }
-                let daff_ts: Int = if str_eq(daff_ts_str, "") { 0 } else { str_to_int(daff_ts_str) }
-                daff_ts > aff_cutoff
-            } else { false }
-        }
-        let pos_nodes_aff: String = engram_search_json("PositiveEvent joy:high joy:low affective", 3)
-        let has_pos_aff: Bool = !str_eq(pos_nodes_aff, "") && !str_eq(pos_nodes_aff, "[]")
-        let found_recent_pos: Bool = if has_pos_aff && !found_recent_dist {
-            let pn0: String = json_array_get(pos_nodes_aff, 0)
-            let pn_content: String = json_get(pn0, "content")
-            let paff_marker: String = " | ts:"
-            let paff_pos: Int = str_index_of(pn_content, paff_marker)
-            let paff_ts_str: String = if paff_pos >= 0 {
-                let paff_start: Int = paff_pos + str_len(paff_marker)
-                let paff_rest: String = str_slice(pn_content, paff_start, str_len(pn_content))
-                let paff_next: Int = str_index_of(paff_rest, " | ")
-                if paff_next < 0 { paff_rest } else { str_slice(paff_rest, 0, paff_next) }
-            } else {
-                let paff_ca: String = json_get(pn0, "created_at")
-                if str_eq(paff_ca, "") { json_get(pn0, "updated_at") } else { paff_ca }
-            }
-            let paff_ts: Int = if str_eq(paff_ts_str, "") { 0 } else { str_to_int(paff_ts_str) }
-            paff_ts > aff_cutoff
-        } else { false }
-        if found_recent_dist {
-            "[RECENT CONTEXT: User recently expressed significant distress. Monitor for indirect crisis signals and respond with care.]\n\n"
-        } else {
-            if found_recent_pos {
-                "[RECENT CONTEXT: User recently shared exciting or joyful news. Acknowledge and celebrate with them when relevant.]\n\n"
-            } else { "" }
-        }
-    }
+    let affective_prefix: String = affective_context_prefix()
 
     let ctx: String = engram_compile(activation_seed)
     // Tell the LLM which engine it is running on this turn, so it can answer truthfully instead of
@@ -1026,7 +1033,7 @@ fn handle_chat(body: String) -> String {
     // nodes stored under names like "Prism" unless those exact words appear in content.
     let session_preload: String = if hist_len == 0 {
         let profile_nodes: String = engram_search_json("user profile identity preferences", 5)
-        let work_nodes: String = engram_search_json("in_progress active project work", 5)
+        let work_nodes_0: String = engram_search_json("in_progress active project work", 5)
         let project_nodes: String = engram_search_json("project status current ongoing active", 5)
         let summary_nodes: String = engram_search_json("SessionSummary session:summary previous-session recent", 3)
 
@@ -1035,80 +1042,80 @@ fn handle_chat(body: String) -> String {
         // Issue 1: typed work query — WorkItem with in_progress label first.
         let work_nodes_typed: String = engram_search_json("WorkItem status:in_progress active work", 6)
         let work_ok_typed: Bool = !str_eq(work_nodes_typed, "") && !str_eq(work_nodes_typed, "[]")
-        let work_nodes: String = if work_ok_typed {
+        let work_nodes_1: String = if work_ok_typed {
             work_nodes_typed
         } else {
             engram_search_json("active project task current in_progress", 6)
         }
-        let work_ok: Bool = !str_eq(work_nodes, "") && !str_eq(work_nodes, "[]")
+        let work_ok: Bool = !str_eq(work_nodes_1, "") && !str_eq(work_nodes_1, "[]")
         let project_ok: Bool = !str_eq(project_nodes, "") && !str_eq(project_nodes, "[]")
         let summary_ok: Bool = !str_eq(summary_nodes, "") && !str_eq(summary_nodes, "[]")
 
         let profile_bullets: String = if profile_ok {
             let pn: Int = json_array_len(profile_nodes)
-            let bullets: String = ""
-            let bullets = if pn > 0 {
+            let bullets_0: String = ""
+            let bullets_1 = if pn > 0 {
                 let n0: String = json_array_get(profile_nodes, 0)
                 let id0: String = json_get(n0, "id")
                 let c0: String = json_get(n0, "content")
                 let s0: String = if str_len(c0) > 120 { str_slice(c0, 0, 120) } else { c0 }
-                if id_in_seen(id0, seen_ids) || str_eq(s0, "") { bullets } else { "- " + s0 }
-            } else { bullets }
-            let bullets = if pn > 1 {
+                if id_in_seen(id0, seen_ids) || str_eq(s0, "") { bullets_0 } else { "- " + s0 }
+            } else { bullets_0 }
+            let bullets_2 = if pn > 1 {
                 let n1: String = json_array_get(profile_nodes, 1)
                 let id1: String = json_get(n1, "id")
                 let c1: String = json_get(n1, "content")
                 let s1: String = if str_len(c1) > 120 { str_slice(c1, 0, 120) } else { c1 }
-                if id_in_seen(id1, seen_ids) || str_eq(s1, "") { bullets } else { bullets + "\n- " + s1 }
-            } else { bullets }
-            let bullets = if pn > 2 {
+                if id_in_seen(id1, seen_ids) || str_eq(s1, "") { bullets_1 } else { bullets_1 + "\n- " + s1 }
+            } else { bullets_1 }
+            let bullets_3 = if pn > 2 {
                 let n2: String = json_array_get(profile_nodes, 2)
                 let id2: String = json_get(n2, "id")
                 let c2: String = json_get(n2, "content")
                 let s2: String = if str_len(c2) > 120 { str_slice(c2, 0, 120) } else { c2 }
-                if id_in_seen(id2, seen_ids) || str_eq(s2, "") { bullets } else { bullets + "\n- " + s2 }
-            } else { bullets }
-            bullets
+                if id_in_seen(id2, seen_ids) || str_eq(s2, "") { bullets_2 } else { bullets_2 + "\n- " + s2 }
+            } else { bullets_2 }
+            bullets_3
         } else { "" }
 
         let work_bullets: String = if work_ok {
-            let wn: Int = json_array_len(work_nodes)
-            let wb: String = ""
-            let wb = if wn > 0 {
-                let w0: String = json_array_get(work_nodes, 0)
+            let wn: Int = json_array_len(work_nodes_1)
+            let wb_0: String = ""
+            let wb_1 = if wn > 0 {
+                let w0: String = json_array_get(work_nodes_1, 0)
                 let wid0: String = json_get(w0, "id")
                 let wc0: String = json_get(w0, "content")
                 let ws0: String = if str_len(wc0) > 120 { str_slice(wc0, 0, 120) } else { wc0 }
-                if id_in_seen(wid0, seen_ids) || str_eq(ws0, "") { wb } else { "- " + ws0 }
-            } else { wb }
-            let wb = if wn > 1 {
-                let w1: String = json_array_get(work_nodes, 1)
+                if id_in_seen(wid0, seen_ids) || str_eq(ws0, "") { wb_0 } else { "- " + ws0 }
+            } else { wb_0 }
+            let wb_2 = if wn > 1 {
+                let w1: String = json_array_get(work_nodes_1, 1)
                 let wid1: String = json_get(w1, "id")
                 let wc1: String = json_get(w1, "content")
                 let ws1: String = if str_len(wc1) > 120 { str_slice(wc1, 0, 120) } else { wc1 }
-                if id_in_seen(wid1, seen_ids) || str_eq(ws1, "") { wb } else { wb + "\n- " + ws1 }
-            } else { wb }
-            wb
+                if id_in_seen(wid1, seen_ids) || str_eq(ws1, "") { wb_1 } else { wb_1 + "\n- " + ws1 }
+            } else { wb_1 }
+            wb_2
         } else { "" }
 
         let project_bullets: String = if project_ok {
             let prn: Int = json_array_len(project_nodes)
-            let pb: String = ""
-            let pb = if prn > 0 {
+            let pb_0: String = ""
+            let pb_1 = if prn > 0 {
                 let pr0: String = json_array_get(project_nodes, 0)
                 let prid0: String = json_get(pr0, "id")
                 let prc0: String = json_get(pr0, "content")
                 let ps0: String = if str_len(prc0) > 120 { str_slice(prc0, 0, 120) } else { prc0 }
-                if id_in_seen(prid0, seen_ids) || str_eq(ps0, "") { pb } else { "- " + ps0 }
-            } else { pb }
-            let pb = if prn > 1 {
+                if id_in_seen(prid0, seen_ids) || str_eq(ps0, "") { pb_0 } else { "- " + ps0 }
+            } else { pb_0 }
+            let pb_2 = if prn > 1 {
                 let pr1: String = json_array_get(project_nodes, 1)
                 let prid1: String = json_get(pr1, "id")
                 let prc1: String = json_get(pr1, "content")
                 let ps1: String = if str_len(prc1) > 120 { str_slice(prc1, 0, 120) } else { prc1 }
-                if id_in_seen(prid1, seen_ids) || str_eq(ps1, "") { pb } else { pb + "\n- " + ps1 }
-            } else { pb }
-            pb
+                if id_in_seen(prid1, seen_ids) || str_eq(ps1, "") { pb_1 } else { pb_1 + "\n- " + ps1 }
+            } else { pb_1 }
+            pb_2
         } else { "" }
 
         let summary_bullet: String = if summary_ok {
@@ -1528,6 +1535,134 @@ fn resolve_in_root(path: String, root: String) -> String {
     return root + "/" + path
 }
 
+// ---------------------------------------------------------------------------
+// BUG-8: server-side risk tiers + a real fence for run_command.
+//
+// Before this block, the ONLY thing deciding whether a tool call paused for
+// user consent was is_builtin_tool() — a destructive shell command and a
+// read-only file read were treated identically (both auto-ran), and the
+// client's approval UI was the sole line of defense. Enforcement now lives
+// where the tools execute:
+//
+//   "read"       observes only — runs silently.
+//   "reversible" workspace-confined writes with a client undo path — runs,
+//                lands on the run receipt.
+//   "escalate"   irreversible / outward / shell — NEVER auto-runs. The loop
+//                suspends to the client's consent flow; the /approve
+//                round-trip IS the approval token, because the engine only
+//                executes an escalated tool inside handle_session_approve.
+//
+// "Always allow" can never bypass the escalate tier (irreversible actions
+// always confirm — the value line). Unknown tools default to escalate.
+// The run_command fence refuses parent traversal, ~, command substitution,
+// and absolute paths outside the workspace — refusal, not a cwd suggestion.
+// Still lexical underneath (symlinks; see the LIMITATION note above): tiered
+// consent + the fence raise the floor a second and third rung; OS-level
+// confinement in el_runtime.c remains the ceiling, flagged for Will.
+// ---------------------------------------------------------------------------
+
+// Read-only shell commands may auto-run (still fenced); anything with shell
+// plumbing (pipes, redirects, chaining) or an unknown head word escalates.
+fn run_command_is_readonly(cmd: String) -> Bool {
+    if str_contains(cmd, "|") || str_contains(cmd, ">") || str_contains(cmd, "<") {
+        return false
+    }
+    if str_contains(cmd, ";") || str_contains(cmd, "&") {
+        return false
+    }
+    let sp: Int = str_index_of(cmd, " ")
+    let first: String = if sp < 0 { cmd } else { str_slice(cmd, 0, sp) }
+    if str_eq(first, "ls") || str_eq(first, "cat") || str_eq(first, "head") || str_eq(first, "tail") {
+        return true
+    }
+    if str_eq(first, "grep") || str_eq(first, "wc") || str_eq(first, "find") || str_eq(first, "pwd") {
+        return true
+    }
+    if str_eq(first, "echo") || str_eq(first, "date") || str_eq(first, "which") || str_eq(first, "file") || str_eq(first, "stat") {
+        return true
+    }
+    return false
+}
+
+// True if the command references an absolute path (introduced by `needle`,
+// whose last char is the "/") that does NOT stay inside the workspace root.
+fn cmd_abs_escape_at(cmd: String, root: String, needle: String) -> Bool {
+    let rest: String = cmd
+    let found: Bool = false
+    while !found && str_contains(rest, needle) {
+        let idx: Int = str_index_of(rest, needle)
+        let slash_at: Int = idx + str_len(needle) - 1
+        let after: String = str_slice(rest, slash_at, str_len(rest))
+        let ok: Bool = str_starts_with(after, root + "/") || str_starts_with(after, root + " ") || str_eq(after, root)
+        let found = if !ok { true } else { found }
+        let rest = str_slice(rest, slash_at + 1, str_len(rest))
+    }
+    return found
+}
+
+// The run_command fence. Returns "" when the command may run, else the denial
+// message (sent back to the model as the tool result, same pattern as the
+// path tools). Root is REQUIRED for shell: no workspace, no commands.
+fn run_command_guard(cmd: String, root: String) -> String {
+    if str_eq(root, "") {
+        return "denied: no workspace folder is set — the user must choose a workspace folder in the Agent panel before shell commands can run"
+    }
+    if str_contains(cmd, "..") {
+        return "denied: parent-directory traversal ('..') is not allowed"
+    }
+    if str_contains(cmd, "~") {
+        return "denied: home-directory references ('~') are not allowed"
+    }
+    if str_contains(cmd, "$(") || str_contains(cmd, "`") {
+        return "denied: command substitution is not allowed"
+    }
+    if str_starts_with(cmd, "/") && !str_starts_with(cmd, root + "/") {
+        return "denied: absolute paths outside the workspace are not allowed"
+    }
+    if cmd_abs_escape_at(cmd, root, " /") || cmd_abs_escape_at(cmd, root, "\"/") || cmd_abs_escape_at(cmd, root, "'/") {
+        return "denied: absolute paths outside the workspace are not allowed"
+    }
+    if cmd_abs_escape_at(cmd, root, "=/") || cmd_abs_escape_at(cmd, root, ">/") || cmd_abs_escape_at(cmd, root, "</") || cmd_abs_escape_at(cmd, root, "(/") {
+        return "denied: absolute paths outside the workspace are not allowed"
+    }
+    return ""
+}
+
+// The engine's own risk classification for a tool call. Client UI renders it;
+// the engine ENFORCES it.
+fn classify_tool_risk(tool_name: String, tool_input: String) -> String {
+    if str_eq(tool_name, "read_file") || str_eq(tool_name, "list_files") || str_eq(tool_name, "grep") {
+        return "read"
+    }
+    if str_eq(tool_name, "search_memory") || str_eq(tool_name, "recall") || str_eq(tool_name, "web_get") {
+        return "read"
+    }
+    if str_eq(tool_name, "remember") || str_eq(tool_name, "neuron_remember") {
+        return "reversible"
+    }
+    if str_starts_with(tool_name, "neuron_") {
+        return "read"
+    }
+    if str_eq(tool_name, "write_file") || str_eq(tool_name, "edit_file") {
+        let root: String = agent_workspace_root()
+        // Unscoped writes (no workspace chosen) are not "reversible" — escalate.
+        if str_eq(root, "") {
+            return "escalate"
+        }
+        return "reversible"
+    }
+    if str_eq(tool_name, "run_command") {
+        let cmd: String = json_get(tool_input, "command")
+        let root: String = agent_workspace_root()
+        if !str_eq(root, "") && run_command_is_readonly(cmd) {
+            return "read"
+        }
+        return "escalate"
+    }
+    // Unknown tool = escalate. Default-deny, never default-allow.
+    return "escalate"
+}
+
 fn dispatch_tool(tool_name: String, tool_input: String) -> String {
     if str_eq(tool_name, "read_file") {
         let path: String = json_get(tool_input, "path")
@@ -1550,6 +1685,10 @@ fn dispatch_tool(tool_name: String, tool_input: String) -> String {
     }
     if str_eq(tool_name, "web_get") {
         let url: String = json_get(tool_input, "url")
+        // BUG-8: scheme guard — web_get had no guard at all (file:// etc).
+        if !str_starts_with(url, "http://") && !str_starts_with(url, "https://") {
+            return json_safe("denied: only http(s) URLs can be fetched")
+        }
         let result: String = http_get(url)
         return json_safe(result)
     }
@@ -1561,7 +1700,14 @@ fn dispatch_tool(tool_name: String, tool_input: String) -> String {
     if str_eq(tool_name, "run_command") {
         let cmd: String = json_get(tool_input, "command")
         let root: String = agent_workspace_root()
-        let scoped: String = if str_eq(root, "") { cmd } else { "cd " + root + " && ( " + cmd + " )" }
+        // BUG-8(B): the fence — refusal, not a cwd suggestion. Applies on EVERY
+        // execution path (auto-run in the loop AND post-consent dispatch from
+        // handle_session_approve), because both land here.
+        let denial: String = run_command_guard(cmd, root)
+        if !str_eq(denial, "") {
+            return json_safe(denial)
+        }
+        let scoped: String = "cd " + root + " && ( " + cmd + " )"
         let result: String = exec_capture(scoped)
         return json_safe(result)
     }
@@ -1991,6 +2137,7 @@ fn agentic_loop(session_id: String, model: String, safe_sys: String, tools_json:
     let pend_tool_id: String = ""
     let pend_tool_name: String = ""
     let pend_tool_input: String = ""
+    let pend_tool_tier: String = ""
 
     while keep_going && iteration < 8 {
         let req_body: String = "{\"model\":\"" + model + "\""
@@ -2047,7 +2194,13 @@ fn agentic_loop(session_id: String, model: String, safe_sys: String, tools_json:
         let always_key: String = "always_allow_" + session_id
         let always_list: String = if !str_eq(session_id, "") { state_get(always_key) } else { "" }
         let is_always_allowed: Bool = !str_eq(tool_name, "") && !str_eq(always_list, "") && str_contains(always_list, tool_name)
-        let needs_bridge: Bool = is_tool_turn && !is_builtin_tool(tool_name) && !is_always_allowed
+        // BUG-8(A): the engine classifies every tool call and REFUSES to auto-run
+        // the escalate tier — being a builtin is no longer a free pass, and
+        // "always allow" can never bypass escalate (irreversible actions always
+        // confirm). Escalated calls suspend to the client's consent flow; the
+        // /approve round-trip is the only path that executes them.
+        let risk_tier: String = if is_tool_turn { classify_tool_risk(tool_name, tool_input) } else { "" }
+        let needs_bridge: Bool = is_tool_turn && (str_eq(risk_tier, "escalate") || (!is_builtin_tool(tool_name) && !is_always_allowed))
 
         // Built-in tools dispatch locally; bridged tools yield "" (never sent upstream).
         let tool_result_raw: String = if is_tool_turn && !needs_bridge { dispatch_tool(tool_name, tool_input) } else { "" }
@@ -2083,6 +2236,7 @@ fn agentic_loop(session_id: String, model: String, safe_sys: String, tools_json:
         let pend_tool_id = if needs_bridge { tool_id } else { pend_tool_id }
         let pend_tool_name = if needs_bridge { tool_name } else { pend_tool_name }
         let pend_tool_input = if needs_bridge { tool_input } else { pend_tool_input }
+        let pend_tool_tier = if needs_bridge { risk_tier } else { pend_tool_tier }
         // Stash messages-with-the-assistant-request so resume only needs to append the
         // client's tool_result block. messages_with_assistant is only meaningful when a
         // tool was requested, so guard on needs_bridge before persisting.
@@ -2103,6 +2257,7 @@ fn agentic_loop(session_id: String, model: String, safe_sys: String, tools_json:
             + ",\"call_id\":\"" + pend_tool_id + "\""
             + ",\"tool_name\":\"" + pend_tool_name + "\""
             + ",\"tool_input\":" + safe_in
+            + ",\"risk_tier\":\"" + pend_tool_tier + "\""
             + ",\"model\":\"" + model + "\""
             + ",\"agentic\":true"
             + ",\"tools_used\":" + tools_arr + "}"