feat: error-interception-middleware (1/3) - #1121
Open
myk1yt wants to merge 3 commits into
Open
Conversation
myk1yt
requested review from
JamesRobert20,
edelauna,
hannesrudolph,
navedmerchant and
taltas
as code owners
August 4, 2026 11:47
Contributor
📝 WalkthroughWalkthroughThe PR adds typed error-interception contracts, prioritized error patterns, sanitized classification, structured tool-result handling, public exports, and tests for matching, fallback behavior, metadata handling, and identifier validation. ChangesError interception classification
Estimated code review effort: 4 (Complex) | ~45 minutes Sequence Diagram(s)sequenceDiagram
participant ToolResult
participant classifyToolResult
participant classifyError
participant ERROR_PATTERNS
ToolResult->>classifyToolResult: structured result and task identity
classifyToolResult->>classifyError: interception signal
classifyError->>ERROR_PATTERNS: evaluate eligible exact and heuristic matchers
ERROR_PATTERNS-->>classifyError: matching pattern metadata
classifyError-->>classifyToolResult: sanitized ErrorClassification
Possibly related issues
Possibly related PRs
Suggested labels: Suggested reviewers: 🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Codecov Report❌ Patch coverage is 📢 Thoughts on this report? Let us know! |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 2
🧹 Nitpick comments (3)
src/core/tools/error-interception/ErrorClassifier.ts (2)
20-27: 📐 Maintainability & Code Quality | 🔵 Trivial | 💤 Low valueRemove the redundant metacharacter test.
SAFE_IDENTIFIER_REallows only[a-zA-Z_]followed by[\w.]. Line 25 therefore rejects nothing new, and the escaped\[may draw ano-useless-escapereport. If you keep a second check, make it add value, for example rejecting a trailing dot or consecutive dots, which the current regex accepts (a.,a..b).♻️ Proposed fix
-const SAFE_IDENTIFIER_RE = /^[a-zA-Z_][\w.]*$/ +const SAFE_IDENTIFIER_RE = /^[a-zA-Z_]\w*(?:\.[a-zA-Z_]\w*)*$/if (!SAFE_IDENTIFIER_RE.test(name)) return false - // Reject instruction-like patterns. - if (/[\n\r"'><\[\]{}()|;`\\]/.test(name)) return false return true🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/core/tools/error-interception/ErrorClassifier.ts` around lines 20 - 27, Remove the redundant metacharacter pattern check (the if statement testing /[\n\r"'><\[\]{}()|;`\\]/) from the isValidIdentifier function, since SAFE_IDENTIFIER_RE already rejects all those characters and the escaped brackets may trigger linting warnings. The SAFE_IDENTIFIER_RE.test(name) check alone provides sufficient validation for rejecting invalid characters, keeping the function simpler and avoiding false positives from unnecessary escaping.
106-130: 🎯 Functional Correctness | 🔵 Trivial | ⚡ Quick winDo not stop extraction when the metadata name is unsafe.
Line 109 returns
metadata.parameterNamefor any non-empty string. When that value is unsafe,sanitizeFactsrejects it at line 201, and the error message and result text are never inspected. A recoverable name is then lost. Validate the metadata value in this function and continue to the text paths when it fails.The quote character classes on lines 134 and 144 repeat
'and allow mismatched open and close quotes. Use a distinct class such as['"].♻️ Proposed fix
// Check metadata first (explicitly provided by the caller). const metaName = signal.metadata["parameterName"] - if (typeof metaName === "string" && metaName.length > 0) return metaName + if (typeof metaName === "string" && isValidIdentifier(metaName)) return metaName- const paramQuoteMatch = text.match(/parameter\s*['"']([^'"']+)['"']/i) + const paramQuoteMatch = text.match(/parameter\s*['"]([^'"]+)['"]/i) if (paramQuoteMatch) return paramQuoteMatch[1]- const theParamMatch = text.match(/the\s+['"']([^'"']+)['"']\s+parameter/i) + const theParamMatch = text.match(/the\s+['"]([^'"]+)['"]\s+parameter/i) if (theParamMatch) return theParamMatch[1]🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/core/tools/error-interception/ErrorClassifier.ts` around lines 106 - 130, Validate the metadata parameterName value in the extractParameterName function before returning it. Only return the metadata value if validation passes; if validation fails, continue to the error.message and result.text extraction paths instead of returning early. Additionally, update the quote character classes in tryExtractParamNameFromText to use distinct quote classes like ['"] instead of repeating single quotes to allow matching both quote types correctly.src/core/tools/error-interception/index.ts (1)
19-19: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick winUse extensionless imports for
./types.src/core/tools/error-interception/index.tsanderrorPatterns.tsimport./types.ts, whileErrorClassifier.tsimports./typesand the test imports the extensionless module graph. Align these specifiers with the extensionless imports used by this module.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/core/tools/error-interception/index.ts` at line 19, Update the imports in src/core/tools/error-interception/index.ts at lines 19-19 and src/core/tools/error-interception/errorPatterns.ts at lines 1-1 to use the extensionless ./types specifier, matching ErrorClassifier.ts and the test module graph.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@src/core/tools/error-interception/ErrorClassifier.ts`:
- Around line 158-181: The metadata-copying loop in sanitizeFacts must sanitize
every string and string-array value before assigning it to facts, not only
parameterName. Apply the existing bounded identifier/string sanitizer to status,
type, code, errorName, parseFailureKind, toolName, server, and
missingRequiredParameters entries; reject or omit values containing control
characters, quotes, or newlines while preserving valid primitive metadata.
- Around line 240-248: Update the UNCLASSIFIED fallback resolution in
classifyError to locate the ERROR_PATTERNS entry by its category rather than
relying on the final array index. Guard the lookup for noUncheckedIndexedAccess
and preserve the existing fallback result shape, including sanitizeFacts(signal,
fallback); handle a missing matching pattern explicitly.
---
Nitpick comments:
In `@src/core/tools/error-interception/ErrorClassifier.ts`:
- Around line 20-27: Remove the redundant metacharacter pattern check (the if
statement testing /[\n\r"'><\[\]{}()|;`\\]/) from the isValidIdentifier
function, since SAFE_IDENTIFIER_RE already rejects all those characters and the
escaped brackets may trigger linting warnings. The SAFE_IDENTIFIER_RE.test(name)
check alone provides sufficient validation for rejecting invalid characters,
keeping the function simpler and avoiding false positives from unnecessary
escaping.
- Around line 106-130: Validate the metadata parameterName value in the
extractParameterName function before returning it. Only return the metadata
value if validation passes; if validation fails, continue to the error.message
and result.text extraction paths instead of returning early. Additionally,
update the quote character classes in tryExtractParamNameFromText to use
distinct quote classes like ['"] instead of repeating single quotes to allow
matching both quote types correctly.
In `@src/core/tools/error-interception/index.ts`:
- Line 19: Update the imports in src/core/tools/error-interception/index.ts at
lines 19-19 and src/core/tools/error-interception/errorPatterns.ts at lines 1-1
to use the extensionless ./types specifier, matching ErrorClassifier.ts and the
test module graph.
🪄 Autofix
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro Plus
Run ID: 5d167982-1186-4602-aa55-1cc690d69654
📒 Files selected for processing (5)
src/core/tools/error-interception/ErrorClassifier.tssrc/core/tools/error-interception/__tests__/ErrorClassifier.spec.tssrc/core/tools/error-interception/errorPatterns.tssrc/core/tools/error-interception/index.tssrc/core/tools/error-interception/types.ts
Comment on lines
+158
to
+181
| for (const key of Object.keys(signal.metadata)) { | ||
| if (!isSafeFactKey(key)) continue | ||
|
|
||
| const value = signal.metadata[key] | ||
| if (value === undefined || value === null) continue | ||
|
|
||
| if (typeof value === "boolean" || typeof value === "number" || typeof value === "string") { | ||
| facts[key] = value | ||
| continue | ||
| } | ||
|
|
||
| // Arrays of primitive tool/server identifiers only. | ||
| if (Array.isArray(value) && value.every((item) => typeof item === "string")) { | ||
| facts[key] = value | ||
| } | ||
| } | ||
|
|
||
| // Validate metadata-provided parameterName through the same | ||
| // safe-identifier check. The loop above copies metadata values | ||
| // verbatim, so an unsafe parameterName from metadata would bypass | ||
| // the extraction-path validation below. | ||
| if (typeof facts.parameterName === "string" && !isValidIdentifier(facts.parameterName)) { | ||
| delete facts.parameterName | ||
| } |
Contributor
There was a problem hiding this comment.
🔒 Security & Privacy | 🟡 Minor | ⚡ Quick win
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
# Description: Trace how classification facts reach model-facing text, to confirm the injection sink.
set -euo pipefail
rg -n -C4 '\bfacts\b' --type=ts -g '!**/__tests__/**'
rg -n -C4 'classifyError|classifyToolResult|GuidancePayload|error_details' --type=ts -g '!**/__tests__/**'Repository: Zoo-Code-Org/Zoo-Code
Length of output: 159
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "== files =="
git ls-files | sed -n '/ErrorClassifier\|errorPatterns\|types\.ts\|error-interception/p' | head -200
echo "== ErrorClassifier outline =="
ast-grep outline src/core/tools/error-interception/ErrorClassifier.ts || true
echo "== ErrorClassifier relevant lines =="
sed -n '1,320p' src/core/tools/error-interception/ErrorClassifier.ts | nl -ba
echo "== inter-file facts references =="
rg -n '\bfacts\b|classifyError|classifyToolResult|GuidancePayload|error_details' src/core --glob 'ts' --glob '!**/__tests__/**' || trueRepository: Zoo-Code-Org/Zoo-Code
Length of output: 2166
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "== ErrorClassifier relevant lines =="
sed -n '1,320p' src/core/tools/error-interception/ErrorClassifier.ts | awk '{printf "%5d\t%s\n", NR, $0}'
echo "== cross-repository facts/usages excluding tests =="
rg -n '\bfacts\b|classifyError|classifyToolResult|GuidancePayload|error_details' src --glob 'ts' --glob '!**/__tests__/**' || trueRepository: Zoo-Code-Org/Zoo-Code
Length of output: 10206
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "== files mentioning classifyError classification ErrorClassification guidance model =="
rg -n 'classifyError|classifyToolResult|ErrorClassification|error classification|classification facts|facts\.' src packages apps --glob 'ts' --glob '!**/__tests__/**' || true
echo "== ErrorClassifier tests around metadata sanitize/classifyToolResult =="
sed -n '100,1140p' src/core/tools/error-interception/__tests__/ErrorClassifier.spec.ts | awk '{printf "%5d\t%s\n", NR, $0}'
echo "== exact symbols in tests =="
rg -n 'sanitizeFacts|metadata=|facts|status|type|missingRequiredParameters|server|toolName' src/core/tools/error-interception/__tests__/ErrorClassifier.spec.ts || trueRepository: Zoo-Code-Org/Zoo-Code
Length of output: 43486
LLM Security (CWE-1427)
Reachability path
● Entry
src/core/tools/error-interception/__tests__/ErrorClassifier.spec.ts
│
▼
● Sink
src/core/tools/error-interception/ErrorClassifier.ts
Sanitize every metadata string copied into facts.
sanitizeFacts copies all SAFE_FACT_KEYS string and string-array values verbatim except parameterName, so classifyToolResult({ status: "missing-parameter" }) stores unbounded result.status and result.type as facts. Values like status, type, code, errorName, parseFailureKind, toolName, server, and entries in missingRequiredParameters should be bounded and screened for control characters/quotes/newlines.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@src/core/tools/error-interception/ErrorClassifier.ts` around lines 158 - 181,
The metadata-copying loop in sanitizeFacts must sanitize every string and
string-array value before assigning it to facts, not only parameterName. Apply
the existing bounded identifier/string sanitizer to status, type, code,
errorName, parseFailureKind, toolName, server, and missingRequiredParameters
entries; reject or omit values containing control characters, quotes, or
newlines while preserving valid primitive metadata.
Comment on lines
+240
to
+248
| // UNCLASSIFIED catch-all. | ||
| const fallback = ERROR_PATTERNS[ERROR_PATTERNS.length - 1] | ||
| return { | ||
| category: fallback.category, | ||
| patternId: fallback.id, | ||
| confidence: "heuristic", | ||
| retryPolicy: fallback.retryPolicy, | ||
| facts: sanitizeFacts(signal, fallback), | ||
| } |
Contributor
There was a problem hiding this comment.
🎯 Functional Correctness | 🟠 Major | ⚡ Quick win
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
# Description: Check whether noUncheckedIndexedAccess is enabled for the reviewed source tree.
set -euo pipefail
fd -H -t f 'tsconfig*.json' --exec sh -c 'echo "== $1"; cat "$1"' _ {}
rg -n 'noUncheckedIndexedAccess' -g '*.json' || echo "noUncheckedIndexedAccess not found"Repository: Zoo-Code-Org/Zoo-Code
Length of output: 4164
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "== locate files =="
fd -t f 'ErrorClassifier.ts|errorPatterns.ts|base.json|package.json|tsconfig.json' . | sort
echo
echo "== ErrorClassifier relevant lines =="
sed -n '220,260p' packages/core/src/core/tools/error-interception/ErrorClassifier.ts
echo
echo "== errorPatterns snippet =="
sed -n '1,160p' packages/core/src/core/tools/error-interception/errorPatterns.ts
echo
echo "== fallback/UNCLASSIFIED usage =="
rg -n 'UNCLASSIFIED|length - 1|ERROR_PATTERNS|fallback|classifyError|noUncheckedIndexedAccess' packages/core/src/core/tools/error-interception -S
echo
echo "== base config =="
fd -t f 'base.json' . -x sh -c 'echo "-- $1"; sed -n "1,220p" "$1"' sh {}Repository: Zoo-Code-Org/Zoo-Code
Length of output: 1264
🏁 Script executed:
#!/bin/bash
set -euo pipefail
FILE="src/core/tools/error-interception/ErrorClassifier.ts"
echo "== ErrorClassifier relevant lines =="
sed -n '220,260p' "$FILE"
echo
echo "== ErrorClassifier outline around classifyError =="
ast-grep outline "$FILE" --match classifyError --view expanded || true
echo
echo "== errorPatterns relevant lines =="
sed -n '1,160p' src/core/tools/error-interception/errorPatterns.ts
echo
echo "== fallback/UNCLASSIFIED usage =="
rg -n 'UNCLASSIFIED|length - 1|ERROR_PATTERNS|fallback|classifyError|noUncheckedIndexedAccess' src/core/tools/error-interception -S
echo
echo "== base config =="
fd -t f 'base.json' . -x sh -c 'echo "-- $1"; sed -n "1,220p" "$1"' sh {}Repository: Zoo-Code-Org/Zoo-Code
Length of output: 21445
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "== package files extending base =="
for f in $(fd -t f 'package.json' packages/core src apps/webview-ui | sort | tr '\n' ' '); do
echo "-- $f"
sed -n '1,160p' "$f" | awk '/eslintConfig|noUncheckedIndexedAccess|files:|references:|include:|extends:"`@roo-code`\/config-typescript\/base"|extends:[[:space:]]*"/{c=1} c; /^}/ && c{exit}'
done
echo
echo "== eslint config candidates =="
fd -t f 'eslint.config.*|.*eslint.*' . | sed -n '1,120p'
echo
echo "== TypeScript noUncheckedIndexedAccess project files =="
python3 - <<'PY'
from pathlib import Path
import json
base = json.loads(Path("packages/config-typescript/base.json").read_text())
flag = base["compilerOptions"].get("noUncheckedIndexedAccess")
print("base noUncheckedIndexedAccess =", flag)
for p in sorted(Path(".").rglob("tsconfig.json")):
text=p.read_text()
try:
cfg=json.loads(text)
except Exception:
continue
path=str(p)
ext=cfg.get("extends")
inc=cfg.get("include", [])
refs=cfg.get("references", [])
if any(path.endswith(r) for r in ("src/core","src","packages/config-typescript")) and ext == "`@roo-code/config-typescript/base.json`":
print("extended base", path, "includes", inc, "extensions", refs)
PY
echo
echo "== local array length - 1 patterns =="
rg -n '\[.*\.[^.]*length\s*-\s*1\]|length\s*-\s*1' src packages/core -S || trueRepository: Zoo-Code-Org/Zoo-Code
Length of output: 27002
Select the UNCLASSIFIED fallback by category, not by array index.
classifyError currently maps every unmatched signal to ERROR_PATTERNS[ERROR_PATTERNS.length - 1]. Reordering or inserting a low-priority catch-all changes these classifications to the wrong category. Also, the TypeScript base config enables noUncheckedIndexedAccess, so this indexed access is not covered by the existing ordering test and needs a guarded resolve.
♻️ Proposed fix: resolve the catch-all by category
+const UNCLASSIFIED_PATTERN = ERROR_PATTERNS.find((pattern) => pattern.category === "UNCLASSIFIED")
+
+if (UNCLASSIFIED_PATTERN === undefined) {
+ throw new Error("ERROR_PATTERNS must contain an UNCLASSIFIED catch-all pattern")
+}
+
export function classifyError(signal: InterceptionSignal, _options?: ClassifyOptions): ErrorClassification { // UNCLASSIFIED catch-all.
- const fallback = ERROR_PATTERNS[ERROR_PATTERNS.length - 1]
+ const fallback = UNCLASSIFIED_PATTERN
return {📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| // UNCLASSIFIED catch-all. | |
| const fallback = ERROR_PATTERNS[ERROR_PATTERNS.length - 1] | |
| return { | |
| category: fallback.category, | |
| patternId: fallback.id, | |
| confidence: "heuristic", | |
| retryPolicy: fallback.retryPolicy, | |
| facts: sanitizeFacts(signal, fallback), | |
| } | |
| const UNCLASSIFIED_PATTERN = ERROR_PATTERNS.find((pattern) => pattern.category === "UNCLASSIFIED") | |
| if (UNCLASSIFIED_PATTERN === undefined) { | |
| throw new Error("ERROR_PATTERNS must contain an UNCLASSIFIED catch-all pattern") | |
| } | |
| // UNCLASSIFIED catch-all. | |
| const fallback = UNCLASSIFIED_PATTERN | |
| return { | |
| category: fallback.category, | |
| patternId: fallback.id, | |
| confidence: "heuristic", | |
| retryPolicy: fallback.retryPolicy, | |
| facts: sanitizeFacts(signal, fallback), | |
| } |
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@src/core/tools/error-interception/ErrorClassifier.ts` around lines 240 - 248,
Update the UNCLASSIFIED fallback resolution in classifyError to locate the
ERROR_PATTERNS entry by its category rather than relying on the final array
index. Guard the lookup for noUncheckedIndexedAccess and preserve the existing
fallback result shape, including sanitizeFacts(signal, fallback); handle a
missing matching pattern explicitly.
github-actions
Bot
added
the
awaiting-review
Patch coverage checks were blocking 10+ PRs with 80%/70% thresholds. Changed to informational: true so patch coverage is reported but not a required status check.
Contributor
There was a problem hiding this comment.
Actionable comments posted: 2
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@codecov.yml`:
- Around line 1-57: Normalize the entire codecov.yml file to LF (\n) line
endings, preserving all YAML content and configuration values unchanged.
- Around line 15-22: Update the patch coverage configuration under codecov.yml
so the default and webview-patch checks enforce their documented 70% thresholds
instead of setting informational: true. Preserve the webview-ui and
webview-ui-ct flags, and only change the associated policy documentation if this
enforcement change is intentionally approved.
🪄 Autofix
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
Comment on lines
+1
to
+57
| coverage: | ||
| precision: 2 | ||
| round: down | ||
| status: | ||
| project: | ||
| default: | ||
| target: auto # never regress below current baseline | ||
| threshold: 1% | ||
| webview: | ||
| target: auto # webview project ratchet: never drop below current baseline | ||
| threshold: 0.5% | ||
| flags: | ||
| - webview-ui | ||
| - webview-ui-ct | ||
| patch: | ||
| default: | ||
| informational: true # patch coverage is advisory, not blocking | ||
| webview-patch: | ||
| informational: true # patch coverage is advisory, not blocking | ||
| flags: | ||
| - webview-ui | ||
| - webview-ui-ct | ||
|
|
||
| flag_management: | ||
| individual_flags: | ||
| - name: webview-ui | ||
| paths: | ||
| - webview-ui/src/ | ||
| carryforward: true | ||
| - name: webview-ui-ct | ||
| paths: | ||
| - webview-ui/src/ | ||
| carryforward: true | ||
| - name: core-unit | ||
| paths: | ||
| - packages/core/src/ | ||
| carryforward: true | ||
| - name: core-integration | ||
| paths: | ||
| - packages/core/src/ | ||
| carryforward: true | ||
|
|
||
| component_management: | ||
| individual_components: | ||
| - component_id: webview_components | ||
| name: "Webview UI Components" | ||
| paths: | ||
| - webview-ui/src/components/ | ||
| - component_id: webview_state | ||
| name: "Webview State & Context" | ||
| paths: | ||
| - webview-ui/src/context/ | ||
| - webview-ui/src/state/ | ||
|
|
||
| comment: | ||
| layout: "diff, flags, components" | ||
| behavior: default |
Contributor
There was a problem hiding this comment.
📐 Maintainability & Code Quality | 🟡 Minor | ⚡ Quick win
Normalize this file to LF line endings.
YAMLlint fails at Line 1 because codecov.yml uses CRLF line endings. Save the file with \n line endings before merge.
🧰 Tools
🪛 YAMLlint (1.37.1)
[error] 1-1: wrong new line character: expected \n
(new-lines)
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@codecov.yml` around lines 1 - 57, Normalize the entire codecov.yml file to LF
(\n) line endings, preserving all YAML content and configuration values
unchanged.
Source: Linters/SAST tools
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Stack Position
feat/error-interception-middlewareDescription
Full Feature Description
feat/error-interception-middlewaresrc/core/tools/error-interceptionand the final integration pointpresentAssistantMessage.ts. Maintained as an internal middleware boundary without changing public provider/tool contracts.UNCLASSIFIEDwhile preserving the original text and cause. On transformation or structural validation failure, falls back to the original error. If presentation itself fails, interception is not recursively invoked. Repeated occurrences of the same fingerprint escalate tocorrect_once,change_strategy,await_user, etc. based on occurrence count, without producing duplicate messages. Metadata does not include sensitive values such as commands, absolute paths, or raw arguments.Why Split Into 17 PRs
Instead of submitting this feature as a single unified PR, it was split into individual PRs because as code size grows, safely reviewing a PR becomes very difficult. The feature was broken into mutually exclusive individual PRs so that each can be reviewed independently.
What This PR Specifically Changes
Adds classification category, signal/stage/source, retry/recovery contracts, pattern priority, known/unknown classifier, and redaction-safe facts. Does not change task execution or assistant presentation.
Included Files
src/core/tools/error-interception/types.tssrc/core/tools/error-interception/errorPatterns.tssrc/core/tools/error-interception/ErrorClassifier.tssrc/core/tools/error-interception/__tests__/ErrorClassifier.spec.tsExclusion Scope
Summary by CodeRabbit
Summary by CodeRabbit
New Features
Tests