Files
coder/coderd/userskills.go
T
Bobby Ho 166d92ba73 fix: bound request body size on JSON API endpoints (#28168)
## Summary

`httpapi.Read` decoded request bodies with no size limit, so a single
request could allocate memory without bound. This adds a 4 MiB default
ceiling, leaves the endpoints that legitimately need more explicitly
exempted, and counts the rejections so a limit set too tight is visible.

This is the first of three PRs split out of #28048, covering the
endpoints that answer in `codersdk.Response` shape. The OAuth2 decode
paths (RFC 6749, RFC 7591) and the SCIM ones (RFC 7644) answer in their
own error shapes and follow in separate PRs, along with the lint rule
that pins the invariant.

Closes PLAT-463. Remediates SEC-416 (CWE-770, CVSS 7.5) and SEC-392.

## Problem

`httpapi.Read` calls `json.NewDecoder(r.Body).Decode(value)` with no
ceiling, and no middleware in the chain bounds body size. The exposure
is pre-authentication: login, OTP, and first-user creation all read a
body before any authorization decision is reached. The existing rate
limiter bounds request *rate*, which is orthogonal to the memory a
single admitted request may consume.

## Fix

`Read` is split into `Read` and `ReadLimit`. `ReadLimit` wraps `r.Body`
in an `http.MaxBytesReader` and keeps the existing decode and validate
logic; `Read` delegates to it with a new `DefaultMaxRequestBodyBytes` of
4 MiB, which covers the 124 remaining non-test callers at a single site.

`http.MaxBytesReader` composes as tightest-wins, so the handlers that
pre-wrapped their own bodies pass their limit to `ReadLimit` rather than
wrapping, and each keeps its previous ceiling byte for byte. That
matters most for the bulk secrets import at `8 * MaxSecretsFileBytes`:
an unconditional wrap inside `Read` would have silently halved it to the
default. `TestImportUserSecretsBodyLargerThanDefaultLimit` is the
regression guard for that specific failure, and
`TestMaxBytesReaderNesting` pins the composition behavior the whole
requirement rests on.

Every rejection site calls `httpapi.RecordRequestBodyLimit`, which names
the limit that tripped on the request's existing log line and marks the
request so `coderd_api_requests_too_large_total{reason="request_body"}`
counts body rejections apart from the 413s coderd answers for other
causes, such as agent log storage overflow. A limit set too tight for a
legitimate payload therefore surfaces without waiting for a user report.

The limit is a constant rather than a deployment option: an operator
raising it to unblock something would reopen the vulnerability as
configuration, where a security scan will not find it. A legitimate 413
is answered with a targeted `ReadLimit` on that endpoint.

## Behavior change

`POST /api/v2/files` now answers 413 rather than 400 when a request body
exceeds `HTTPFileMaxBytes`. It installed that bound already but reported
the rejection as a read failure, which leaked the stdlib `http: request
body too large` string through `Detail` and kept the largest limit in
the tree off the metric. The separate 413 for an oversized expanded
archive is unchanged.

The task log snapshot endpoint now answers 413 rather than 400 when its
64 KiB cap is exceeded. Routing it through `ReadLimit` also changes its
decode-failure message from "Failed to decode request payload." to
"Request body must be valid JSON.", which is what every other endpoint
answers. Its tests are updated to match both.

`coderd_api_requests_too_large_total` is new, so there is no existing
query to migrate. It counts the 413s coderd answers, labeled `method`,
`path`, and `reason`. `reason="request_body"` is a rejection by one of
the limits above; `reason="other"` is a 413 that has nothing to do with
body size, such as agent log storage overflow.

## Reading this

The commits are ordered to be read in sequence. Commits 1 and 2 are the
security fix; commits 3 to 5 are the observability consequences, and
commit 3 is the one that touches dashboards. Commit 7 documents the
limit on the REST API reference index. Commits 6 and 8 add and revert an
exhaustive `@Failure 413` annotation pass, which buried the fix under
its regenerated swagger, and cancel out.
2026-08-18 12:54:45 -07:00

351 lines
10 KiB
Go

package coderd
import (
"context"
"errors"
"fmt"
"net/http"
"github.com/go-chi/chi/v5"
"github.com/google/uuid"
"golang.org/x/xerrors"
"github.com/coder/coder/v2/coderd/audit"
"github.com/coder/coder/v2/coderd/database"
"github.com/coder/coder/v2/coderd/database/db2sdk"
"github.com/coder/coder/v2/coderd/httpapi"
"github.com/coder/coder/v2/coderd/httpmw"
"github.com/coder/coder/v2/coderd/x/skills"
"github.com/coder/coder/v2/codersdk"
)
const (
// personalSkillJSONEscapeExpansion is the maximum expansion for one byte in a JSON string.
personalSkillJSONEscapeExpansion = 6
// personalSkillRequestEnvelopeBytes leaves room for the surrounding JSON object.
personalSkillRequestEnvelopeBytes = 1024
// maxPersonalSkillRequestBytes allows worst-case JSON string escaping for
// otherwise valid raw skill content.
maxPersonalSkillRequestBytes = skills.MaxPersonalSkillSizeBytes*personalSkillJSONEscapeExpansion + personalSkillRequestEnvelopeBytes
// These names are raised by trigger functions with USING CONSTRAINT.
// They are not table CHECK constraints, so dbgen does not emit them in
// check_constraint.go.
userSkillsPerUserLimitConstraint database.CheckConstraint = "user_skills_per_user_limit"
userSkillUserDeletedConstraint database.CheckConstraint = "user_skill_user_deleted"
)
// @Summary Create a user skill
// @ID create-a-user-skill
// @Security CoderSessionToken
// @Accept json
// @Produce json
// @Tags Users
// @Param user path string true "User ID, username, or me"
// @Param request body codersdk.CreateUserSkillRequest true "Create user skill request"
// @Success 201 {object} codersdk.UserSkill
// @Router /api/experimental/users/{user}/skills [post]
// @x-apidocgen {"skip": true}
func (api *API) postUserSkill(rw http.ResponseWriter, r *http.Request) {
var (
ctx = r.Context()
user = httpmw.UserParam(r)
auditor = api.Auditor.Load()
aReq, commitAudit = audit.InitRequest[database.UserSkill](rw, &audit.RequestParams{
Audit: *auditor,
Log: api.Logger,
Request: r,
Action: database.AuditActionCreate,
})
)
defer commitAudit()
var req codersdk.CreateUserSkillRequest
if !httpapi.ReadLimit(ctx, rw, r, maxPersonalSkillRequestBytes, &req) {
return
}
parsedSkill, err := skills.ParsePersonalSkillMarkdown([]byte(req.Content))
if err != nil {
writeInvalidUserSkillContent(ctx, rw, err)
return
}
params := database.InsertUserSkillParams{
ID: uuid.New(),
UserID: user.ID,
Name: parsedSkill.Name,
Description: parsedSkill.Description,
Content: req.Content,
}
skill, err := api.Database.InsertUserSkill(ctx, params)
if err != nil {
if httpapi.IsUnauthorizedError(err) {
httpapi.Forbidden(rw)
return
}
if database.IsCheckViolation(err, userSkillUserDeletedConstraint) {
writeCannotCreateUserSkillForDeletedUser(ctx, rw)
return
}
if httpapi.Is404Error(err) {
httpapi.ResourceNotFound(rw)
return
}
if database.IsCheckViolation(err, userSkillsPerUserLimitConstraint) {
writeUserSkillLimitReached(ctx, rw)
return
}
if database.IsUniqueViolation(err, database.UniqueUserSkillsUserIDNameIndex) {
httpapi.Write(ctx, rw, http.StatusConflict, codersdk.Response{
Message: "A skill with that name already exists.",
Detail: err.Error(),
})
return
}
httpapi.InternalServerError(rw, err)
return
}
aReq.New = skill
httpapi.Write(ctx, rw, http.StatusCreated, db2sdk.UserSkill(skill))
}
// @Summary List user skills
// @ID list-user-skills
// @Security CoderSessionToken
// @Produce json
// @Tags Users
// @Param user path string true "User ID, username, or me"
// @Success 200 {array} codersdk.UserSkillMetadata
// @Router /api/experimental/users/{user}/skills [get]
// @x-apidocgen {"skip": true}
func (api *API) getUserSkills(rw http.ResponseWriter, r *http.Request) { //nolint:revive // Method name matches route.
ctx := r.Context()
user := httpmw.UserParam(r)
rows, err := api.Database.ListUserSkillMetadataByUserID(ctx, user.ID)
if err != nil {
if httpapi.Is404Error(err) {
httpapi.ResourceNotFound(rw)
return
}
httpapi.InternalServerError(rw, err)
return
}
httpapi.Write(ctx, rw, http.StatusOK, db2sdk.UserSkillMetadataList(rows))
}
// @Summary Get a user skill by name
// @ID get-a-user-skill-by-name
// @Security CoderSessionToken
// @Produce json
// @Tags Users
// @Param user path string true "User ID, username, or me"
// @Param skillName path string true "Skill name"
// @Success 200 {object} codersdk.UserSkill
// @Router /api/experimental/users/{user}/skills/{skillName} [get]
// @x-apidocgen {"skip": true}
func (api *API) getUserSkill(rw http.ResponseWriter, r *http.Request) { //nolint:revive // Method name matches route.
ctx := r.Context()
user := httpmw.UserParam(r)
name := chi.URLParam(r, "skillName")
skill, err := api.Database.GetUserSkillByUserIDAndName(ctx, database.GetUserSkillByUserIDAndNameParams{
UserID: user.ID,
Name: name,
})
if err != nil {
if httpapi.Is404Error(err) {
httpapi.ResourceNotFound(rw)
return
}
httpapi.InternalServerError(rw, err)
return
}
httpapi.Write(ctx, rw, http.StatusOK, db2sdk.UserSkill(skill))
}
// @Summary Update a user skill
// @ID update-a-user-skill
// @Security CoderSessionToken
// @Accept json
// @Produce json
// @Tags Users
// @Param user path string true "User ID, username, or me"
// @Param skillName path string true "Skill name"
// @Param request body codersdk.UpdateUserSkillRequest true "Update user skill request"
// @Success 200 {object} codersdk.UserSkill
// @Router /api/experimental/users/{user}/skills/{skillName} [patch]
// @x-apidocgen {"skip": true}
func (api *API) patchUserSkill(rw http.ResponseWriter, r *http.Request) {
var (
ctx = r.Context()
user = httpmw.UserParam(r)
name = chi.URLParam(r, "skillName")
auditor = api.Auditor.Load()
aReq, commitAudit = audit.InitRequest[database.UserSkill](rw, &audit.RequestParams{
Audit: *auditor,
Log: api.Logger,
Request: r,
Action: database.AuditActionWrite,
})
)
defer commitAudit()
var req codersdk.UpdateUserSkillRequest
if !httpapi.ReadLimit(ctx, rw, r, maxPersonalSkillRequestBytes, &req) {
return
}
parsedSkill, err := skills.ParsePersonalSkillMarkdown([]byte(req.Content))
if err != nil {
writeInvalidUserSkillContent(ctx, rw, err)
return
}
if parsedSkill.Name != name {
httpapi.Write(ctx, rw, http.StatusBadRequest, codersdk.Response{
Message: "Skill name in path does not match frontmatter name.",
Detail: fmt.Sprintf("path has %q, frontmatter has %q", name, parsedSkill.Name),
})
return
}
params := database.UpdateUserSkillByUserIDAndNameParams{
UserID: user.ID,
Name: name,
Description: parsedSkill.Description,
Content: req.Content,
}
var (
skill database.UserSkill
oldSkill database.UserSkill
)
err = api.Database.InTx(func(tx database.Store) error {
fetched, err := tx.GetUserSkillByUserIDAndName(ctx, database.GetUserSkillByUserIDAndNameParams{
UserID: user.ID,
Name: name,
})
if err != nil {
return xerrors.Errorf("fetch user skill: %w", err)
}
updated, err := tx.UpdateUserSkillByUserIDAndName(ctx, params)
if err != nil {
return xerrors.Errorf("update user skill: %w", err)
}
oldSkill = fetched
skill = updated
return nil
}, nil)
if err != nil {
if httpapi.IsUnauthorizedError(err) {
httpapi.Forbidden(rw)
return
}
if database.IsCheckViolation(err, userSkillUserDeletedConstraint) {
writeCannotModifyUserSkillForDeletedUser(ctx, rw)
return
}
if httpapi.Is404Error(err) {
httpapi.ResourceNotFound(rw)
return
}
httpapi.InternalServerError(rw, err)
return
}
// Assign audit state after InTx returns so the audit log can never
// claim a rolled-back update was committed.
aReq.Old = oldSkill
aReq.New = skill
httpapi.Write(ctx, rw, http.StatusOK, db2sdk.UserSkill(skill))
}
// @Summary Delete a user skill
// @ID delete-a-user-skill
// @Security CoderSessionToken
// @Tags Users
// @Param user path string true "User ID, username, or me"
// @Param skillName path string true "Skill name"
// @Success 204
// @Router /api/experimental/users/{user}/skills/{skillName} [delete]
// @x-apidocgen {"skip": true}
func (api *API) deleteUserSkill(rw http.ResponseWriter, r *http.Request) {
var (
ctx = r.Context()
user = httpmw.UserParam(r)
name = chi.URLParam(r, "skillName")
auditor = api.Auditor.Load()
aReq, commitAudit = audit.InitRequest[database.UserSkill](rw, &audit.RequestParams{
Audit: *auditor,
Log: api.Logger,
Request: r,
Action: database.AuditActionDelete,
})
)
defer commitAudit()
deleted, err := api.Database.DeleteUserSkillByUserIDAndName(ctx, database.DeleteUserSkillByUserIDAndNameParams{
UserID: user.ID,
Name: name,
})
if err != nil {
if httpapi.Is404Error(err) {
httpapi.ResourceNotFound(rw)
return
}
httpapi.InternalServerError(rw, err)
return
}
aReq.Old = deleted
rw.WriteHeader(http.StatusNoContent)
}
func writeCannotCreateUserSkillForDeletedUser(ctx context.Context, rw http.ResponseWriter) {
httpapi.Write(ctx, rw, http.StatusConflict, codersdk.Response{
Message: "Cannot create skills for deleted users.",
Detail: "This user has been deleted and cannot be modified.",
})
}
func writeCannotModifyUserSkillForDeletedUser(ctx context.Context, rw http.ResponseWriter) {
httpapi.Write(ctx, rw, http.StatusConflict, codersdk.Response{
Message: "Cannot modify skills for deleted users.",
Detail: "This user has been deleted and cannot be modified.",
})
}
func writeUserSkillLimitReached(ctx context.Context, rw http.ResponseWriter) {
httpapi.Write(ctx, rw, http.StatusConflict, codersdk.Response{
Message: "Personal skill limit reached.",
Detail: fmt.Sprintf(
"Each user can have at most %d personal skills.",
skills.MaxPersonalSkillsPerUser,
),
})
}
func writeInvalidUserSkillContent(ctx context.Context, rw http.ResponseWriter, err error) {
message := "Invalid skill content."
switch {
case errors.Is(err, skills.ErrInvalidSkillName):
message = "Invalid skill name."
case errors.Is(err, skills.ErrSkillBodyRequired):
message = "Skill body is required."
case errors.Is(err, skills.ErrSkillTooLarge):
message = "Skill content is too large."
case errors.Is(err, skills.ErrSkillDescriptionTooLarge):
message = "Skill description is too large."
}
httpapi.Write(ctx, rw, http.StatusBadRequest, codersdk.Response{
Message: message,
Detail: err.Error(),
})
}