Three limit problems in the find and ls tools.
**1. `limit: 0` and other values below 1.** find passed `limit` to fd
unchanged. fd treats `--max-results 0` as no limit, so `find({ pattern:
"*", limit: 0 })` returned every match, followed by `[0 results limit
reached. Use limit=0 for more, or refine pattern]`. ls with `limit: 0`
listed nothing and printed `(empty directory)`. A fractional limit such
as `1.5` printed `1.5 results limit reached. Use limit=3`.
Fix: both tools use `Math.max(1, Math.floor(limit ?? DEFAULT_LIMIT))`.
grep already clamps to at least 1. The fd path and the custom `glob`
path share this value.
**2. Notice at exactly N results (#7121).** With 3 matching files and
`limit: 3`, fd returned 3 results. find checked `results >= limit` and
printed `[3 results limit reached. Use limit=6 for more, or refine
pattern]`, even though there were no more files.
Fix: find asks fd for `limit + 1` results, shows the first `limit`, and
prints the notice only when the extra result exists. ls already worked
this way. Custom `FindOperations.glob` now also gets `limit + 1`. The
tool shows at most `limit` of the returned results and prints the notice
only when more were returned. The interface is unchanged. This follows
closed PR #7122 without its unrelated changes.
**3. Results past the 50KB cap.** find kept every fd line in memory.
`truncateHead` then dropped everything past 50KB. find now handles each
line as it arrives. It keeps lines up to the first one past the cap and
only counts the rest, as grep does in #11705. Because that first line
past the cap is kept, `truncateHead` makes the same truncation decision
as before. `details.truncation.totalLines` and `totalBytes` come from
the counts, so the notices and `details` are unchanged.
Behavior differences:
- A custom `glob` that ignores `options.limit` and returns more results
used to have all of them shown. Now the tool shows the first `limit`
results with the limit notice.
- Custom `glob` implementations get `limit + 1` instead of `limit`.
- In the fd path, output made only of blank lines with exit code 0 now
gives `No files found matching pattern` instead of empty text. fd does
not produce such output.
**Testing**
- New tests in `test/tools.test.ts`:
- find: exactly N matches gives no notice, and N+1 matches give the
notice with N lines (#7121). Fails on main.
- find: `limit` 0, -5, 0.5 and 1.9 each show 1 result with the `1
results` notice. Fails on main.
- find with custom `glob`: the operation gets `limit + 1`, and exact and
over-limit results give the right output. Fails on main.
- find past the byte cap: 300 names of 200 bytes each.
`details.truncation` equals `truncateHead` of the full output, except
for `content`. Passes on main too; it guards the counting. It fails if
the `totalLines` override is removed.
- ls: `limit` 0, -5, 0.5 and 1.9 each list 1 entry with the notice.
Fails on main.
- ls: exactly N entries gives no notice. Passes on main too; it guards
the existing behavior.
- `test/tools.test.ts` and the find regression suites 3302, 3303 and
6104 pass (110 tests). `npm run check` passes.
fixes#7121
This PR is AI-generated.
When `context` is set, the grep tool read every matched file whole and
split it into lines, only to print the lines around each match. Memory
grew with the size of the matched files, not the output. In the issue's
reproduction (8 log files of about 95 MB, one match each, `context: 2`,
`--max-old-space-size=1024`), the process ran out of heap after about 10
s.
ripgrep already supports context. This change passes `--context N` to rg
and formats the `context` events from its JSON output. The tool no
longer opens matched files.
Details:
- After the last allowed match (`limit`), rg still sends that match's
trailing context. The tool keeps those lines, then stops rg. A match
inside that window is shown as a context line, as before.
- Lines past the 50 KB output cap are counted but not stored, since
truncation drops them anyway. Match counts, the match-limit notice, and
`details.matchLimitReached` stay as before.
- rg merges overlapping context, so lines shared by nearby matches are
printed once instead of repeated per match. This is the one intended
output change.
- Context lines are decoded by rg, so UTF-16 files with a BOM show
correctly with context (#11423, closed as no-action). Non-UTF-8 lines
and paths, which rg sends base64-encoded, are decoded too.
Two related grep fixes in the same code:
- **Exact match limit:** with exactly N matches and `limit: N`, grep
said "N matches limit reached. Use limit=2N". It now sets the notice and
`details.matchLimitReached` only after rg sends one more match. Cost:
when no further match exists, rg finishes the search instead of being
stopped at the Nth match, the same as a search with fewer than N
matches.
- **Lone CR:** grep removed every `\r` in a line, so `one\rtwo needle`
showed as `onetwo needle` and text copied from grep failed to match in
edit. It now strips only the line terminator (`\n` or `\r\n`), like the
read tool. The TUI still hides CRs when rendering. This updates the
expectation in the #11398 test from `beforeprogress` to
`before\rprogress`.
Breaking: `GrepOperations.readFile` is removed because nothing uses it
anymore. rg always ran locally, so this hook never made grep work on a
remote backend. Docs now say to override `execute()` for remote search,
as the Gondolin example does. Code that passes `readFile` still runs,
and the value is ignored. Only code that type-checks an object literal
against `GrepOperations` fails to compile.
Known remaining differences and limits:
- Memory no longer grows with file size or match count, but a single
huge matching line (tens of MB) is still buffered whole before being cut
to 500 characters. That was already true for `context: 0` on main.
Bounding it would mean discarding oversized rg JSON records, which is a
separate change.
- In a directory search, rg stops reading a file once it finds a NUL
byte, so context past that point in a text-looking file that turns
binary later is omitted. main reread the file and printed it. Explicit
file paths are unaffected.
- `details.truncation.totalLines`/`totalBytes` count formatted rows.
With deduplicated context, or file names containing newlines, they can
differ from main. No renderer or notice displays them.
**Testing**
- Issue reproduction: before, `FATAL ERROR ... JavaScript heap out of
memory`. After, correct output in about 60 ms with 135 MB RSS.
- `pattern: "line", context: 3, limit: 100000` over the same 770 MB of
logs reports both limit notices and peaks at 154 MB RSS.
- New tests in `test/tools.test.ts`:
- Merged overlapping context, plus trailing context at the limit
including a match inside it. Fails on main.
- UTF-16 with context. Fails on main.
- Truncation reported when stored output ends exactly at 50 KB. Guards
the counting logic.
- Matches still counted past the byte cap. Passes on main too; it guards
against stopping rg early at the cap.
- Huge `context` value accepted.
- No limit notice with exactly `limit` matches (with and without
context), notice with more. Fails before this change.
- Lone CR kept and CRLF stripped, with and without context. Fails before
this change.
- All 94 tests in the file pass. `npm run check` passes.
fixes#9276
This PR is AI-generated.
When `context` is set, the grep tool reads the file again to print the
lines around each match. It turned CRLF into LF and also turned every
lone CR into LF before splitting the text into lines. ripgrep, which
supplies the match line numbers, counts lines only at LF. A lone CR
therefore added an extra line on pi's side, so every later line was off
by one.
For example, in a file containing `before\rprogress\nneedle\nafter`, a
search for `needle` with `context: 1` showed `progress` as the line-2
match and left out `after`.
The fix keeps the CRLF normalization and splits only on LF, so pi
numbers lines the same way ripgrep does. Displayed lines already have
`\r` removed, so the line before the match now shows as
`beforeprogress`.
**Testing**
- Added a test in `packages/coding-agent/test/tools.test.ts` for a file
with a lone CR searched with `context: 1`. It fails without the fix and
passes with it.
- `test/tools.test.ts`: all 87 tests pass.
- `npm run check` passes.
- Reproduction from the issue. Before the fix:
```
f.txt-1- before
f.txt:2: progress
f.txt-3- needle
```
After the fix:
```
f.txt-1- beforeprogress
f.txt:2: needle
f.txt-3- after
```
Fixes#11398
Requested by @badlogic
Co-authored-by: triagent <triagent@localhost>
The built-in grep tool spawned rg with the inherited environment, so
RIPGREP_CONFIG_PATH (e.g. a config containing --fixed-strings) silently
changed tool semantics: literal: false searches became literal and the
agent got incomplete results with no error.
Pass --no-config so the tool contract is deterministic. Users who want
custom rg behavior can use the bash tool or wrap grep in an extension.
closes#11203
Structured bash results now hold the full output up to 1 MiB instead of the
model-facing 2000 lines or 50KB, keeping the first and last 512 KiB of longer
output, and add truncated and full_output_path.
This adds codemode and MCP support to pi. Codemode (packages/codemode) runs
model-written JavaScript in a QuickJS wasm VM inside a worker. From there the
script calls pi's tools as async functions, uses a per-session store, and can
read the model catalog and run classifiers. The codemode tool in coding-agent
is compatible with Codex's exec tool, and codemode.mode works like Codex's
tool modes (on or only). Large results are written to a JSON temp file. MCP
(packages/mcp) is a standalone client for stdio and streamable HTTP servers,
with OAuth.
The core only gets general mechanisms; codemode, tool_search and MCP are
built-in extensions that use them. Other extensions can replace these
built-ins. Tools declare an exposure: direct, model-only, codemode, deferred
or hidden. prepareLoadout() lets an orchestrating tool change which tool
declarations the model sees. ctx.executeTool() runs nested calls through
validation and the tool_call/tool_result hooks. Those calls emit events with
parentToolCallId and are recorded as bounded nestedCalls on the parent
result, which compaction and HTML export use. Tools can also return
structuredContent with an outputSchema, and can report errors with isError
instead of throwing. MCP servers come from mcp.json (global, or per project
after the project is trusted) or from pi.registerMcpServer(). They are
managed with /mcp and pi mcp list|login|logout, and their runtimes load on
first use.
Closes#10040
Map signal termination to conventional nonzero exit codes and reject null exit codes from custom shell operations while preserving partial output.
fixes#9577
When tools are registered by an extension (e.g. Kimchi) the execute callback
receives an ExtensionContext with the real session cwd. Use ctx.cwd || cwd so
tools resolve paths against the current session cwd instead of the cwd captured
at tool creation time.
Applies to read, write, edit, grep, find, ls, bash, and powershell.
Co-authored-by: Kimchi <noreply@kimchi.dev>
* fix(coding-agent): report edit access failures correctly closes#3894
- classify edit and edit-preview access errors by errno
- add regressions for missing, permission, and fallback cases
- document the fix in the coding-agent changelog
* chore: get rid of CHANGELOG.md entry
* refactor(coding-agent): apply review suggestions - use single error msg
* refactor(coding-agent): clean up test cases
When 429/500 errors occur during tool execution, empty assistant messages
with stopReason='error' get persisted. These break the tool_use -> tool_result
chain for Claude/Gemini APIs.
Added centralized filtering in transformMessages to skip assistant messages
with empty content and no tool calls. Provider-level filters remain for
defense-in-depth.
- Validate working directory exists before spawning to provide clear error message
- Add spawn error handler to prevent uncaught exceptions when shell not found or cwd invalid
- Add tests for both error scenarios
Without these fixes, spawn errors (e.g., ENOENT from missing cwd or shell) would
cause uncaught exceptions that crash the entire agent session instead of being
returned as clean tool errors.
Co-authored-by: robinwander <robinwander@users.noreply.github.com>
Add grep, find, and ls tools for safe code exploration without modification risk.
These tools are available via the new --tools CLI flag.
- grep: Uses ripgrep (auto-downloaded) for fast regex searching. Respects .gitignore,
supports glob filtering, context lines, and hidden files.
- find: Uses fd (auto-downloaded) for fast file finding. Respects .gitignore, supports
glob patterns, and hidden files.
- ls: Lists directory contents with proper sorting and directory indicators.
- --tools flag: Specify available tools (e.g., --tools read,grep,find,ls for read-only mode)
- Dynamic system prompt adapts to selected tools with relevant guidelines
Closes#74
- Add generateDiffString() function in edit tool to create unified diffs with line numbers and 4 lines of context
- Store only the formatted diff string in tool result details instead of full file contents
- Update tool-execution renderer to parse and colorize the diff string
- Filter out message_update events from session saving to prevent verbose session files
- Add markdown nested list and table rendering tests
**Improved output messages:**
1. File fits within limits: Just outputs content (no notices)
2. Lines get truncated: Shows "Some lines were truncated to 2000 characters for display"
3. File doesn't fit limit: Shows "N more lines not shown. Use offset=X to continue reading"
4. Offset beyond file: Shows "Error: Offset X is beyond end of file (N lines total)"
5. Both truncations: Combines both notices with ". " separator
**Comprehensive test coverage:**
- Files within limits (no notices)
- Large files (line truncation)
- Long lines (character truncation)
- Offset parameter
- Limit parameter
- Offset + limit together
- Invalid offset (out of bounds)
- Combined truncations (both notices)
All 17 tests passing ✓
Tool results now use content blocks and can include both text and images.
All providers (Anthropic, Google, OpenAI Completions, OpenAI Responses)
correctly pass images from tool results to LLMs.
- Update ToolResultMessage type to use content blocks
- Add placeholder text for image-only tool results in Google/Anthropic
- OpenAI providers send tool result + follow-up user message with images
- Fix Anthropic JSON parsing for empty tool arguments
- Add comprehensive tests for image-only and text+image tool results
- Update README with tool result content blocks API
