slopguard_core/rule/
mod.rs

1use std::collections::HashSet;
2use std::fmt;
3use std::fs::read_to_string;
4use std::io::Error as IoError;
5use std::path::{Component, Path, PathBuf};
6use std::str::{from_utf8, Utf8Error};
7
8use ast_grep_language::SupportLang;
9use ignore::WalkBuilder;
10use serde::{Deserialize, Serialize};
11use serde_yaml::Value;
12use slopguard_rules::BuiltinRules;
13use strum::{Display, EnumIter, EnumString};
14use thiserror::Error;
15
16use crate::cross_file::CrossFileKind;
17use crate::metric::Metric;
18
19#[derive(Debug, Clone, PartialEq, Eq, Hash, Serialize, Deserialize)]
20pub struct RuleId(pub String);
21
22impl fmt::Display for RuleId {
23    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
24        f.write_str(&self.0)
25    }
26}
27
28impl RuleId {
29    pub fn as_str(&self) -> &str {
30        &self.0
31    }
32}
33
34impl From<String> for RuleId {
35    fn from(s: String) -> Self {
36        Self(s)
37    }
38}
39
40impl From<&str> for RuleId {
41    fn from(s: &str) -> Self {
42        Self(s.to_string())
43    }
44}
45
46#[derive(Debug, Display, Clone, PartialEq, Eq, Serialize, Deserialize)]
47#[serde(rename_all = "lowercase")]
48#[strum(serialize_all = "lowercase")]
49pub enum Severity {
50    Error,
51    Warning,
52}
53
54#[derive(Debug, Display, EnumIter, Clone, PartialEq, Eq, Hash, Serialize, Deserialize)]
55#[serde(rename_all = "lowercase")]
56#[strum(serialize_all = "lowercase")]
57pub enum Language {
58    Rust,
59    TypeScript,
60}
61
62impl Language {
63    /// The ast-grep languages a rule written for this language is compiled
64    /// against. TypeScript rules also cover TSX files.
65    pub fn ast_grep_langs(&self) -> &'static [SupportLang] {
66        match self {
67            Language::Rust => &[SupportLang::Rust],
68            Language::TypeScript => &[SupportLang::TypeScript, SupportLang::Tsx],
69        }
70    }
71}
72
73#[derive(Debug, Display, EnumString, Default, Clone, PartialEq, Eq, Serialize, Deserialize)]
74#[serde(rename_all = "lowercase")]
75#[strum(serialize_all = "lowercase")]
76pub enum Category {
77    Slop,
78    Security,
79    #[default]
80    Correctness,
81}
82
83impl Category {
84    /// Derive the category from the first directory of a rule path relative to
85    /// its ruleset root (e.g. `security/no-debug.yml`). Returns `None` when the
86    /// path has no directory or the directory is not a known category.
87    fn from_rule_path(path: &Path) -> Option<Self> {
88        match path.components().next()? {
89            Component::Normal(dir) => dir.to_str()?.parse().ok(),
90            _ => None,
91        }
92    }
93}
94
95#[derive(Debug, Clone, Serialize, Deserialize)]
96pub struct RuleTests {
97    #[serde(default)]
98    pub should_match: Vec<String>,
99    #[serde(default)]
100    pub should_not_match: Vec<String>,
101    /// Whole-file fixtures for metric rules, resolved against the embedded
102    /// ruleset root (builtin rules) or the custom rule directory the rule was
103    /// loaded from. A metric cannot be measured on a snippet.
104    #[serde(default)]
105    pub should_match_files: Vec<String>,
106    #[serde(default)]
107    pub should_not_match_files: Vec<String>,
108}
109
110/// How the `reason` (the finding's `note`) is produced when an `ai_check` rule
111/// fires. See [`AiCheck::reason`].
112#[derive(Debug, Display, Clone, Copy, PartialEq, Eq, Serialize, Deserialize, Default)]
113#[serde(rename_all = "lowercase")]
114#[strum(serialize_all = "lowercase")]
115pub enum ReasonMode {
116    /// Use the rule's static `note` as the reason. No generative LLM call.
117    #[default]
118    Static,
119    /// Escalate to the generative LLM to write a per-instance reason. Reserved
120    /// for relational rules whose message depends on the specific match.
121    Generated,
122}
123
124/// Optional AI confirmation step for a rule. When present, the rule's `rule`
125/// AST pattern acts as a cheap pre-filter: matches become candidates that an
126/// LLM (or the System One classifier) must confirm before they are reported.
127/// The AST layer never executes this field; the AI pipeline in `slopguard-ai`
128/// reads it.
129#[derive(Debug, Clone, PartialEq, Serialize, Deserialize)]
130pub struct AiCheck {
131    /// Prompt template. Double use: the generative LLM prompt (variables
132    /// `{{code}}`, `{{filename}}`, `{{rule_context}}`), and the `instructions`
133    /// of the System One classifier's noul question.
134    pub prompt: String,
135    /// Model override for this rule. Falls back to `[ai].model` in config.
136    #[serde(default)]
137    pub model: Option<String>,
138    /// How the finding's reason is produced: a `static` note (default) or a
139    /// per-instance `generated` note written by the LLM.
140    #[serde(default)]
141    pub reason: ReasonMode,
142    /// Per-rule classifier probability threshold, overriding
143    /// `[ai.classifier].threshold`. Fires when `p >= threshold`.
144    #[serde(default)]
145    pub threshold: Option<f64>,
146    /// Classifier calibration: what a "yes" (issue) looks like. Maps to the
147    /// noul question's `criteria.true`.
148    #[serde(default)]
149    pub if_true: Option<String>,
150    /// Classifier calibration: what a "no" (not an issue) looks like. Maps to
151    /// the noul question's `criteria.false`.
152    #[serde(default)]
153    pub if_false: Option<String>,
154}
155
156fn default_enabled() -> bool {
157    true
158}
159
160#[derive(Debug, Clone, Serialize, Deserialize)]
161pub struct Rule {
162    pub id: RuleId,
163    pub language: Language,
164    pub severity: Severity,
165    pub message: String,
166    #[serde(default)]
167    pub note: Option<String>,
168    #[serde(default)]
169    pub category: Option<Category>,
170    #[serde(default)]
171    pub fix: Option<String>,
172    #[serde(default)]
173    pub rule: Value,
174    /// File-level metric evaluated instead of `rule`. Mutually exclusive with it.
175    #[serde(default)]
176    pub metric: Option<Metric>,
177    /// The value `metric` must exceed for the rule to fire. Required with `metric`.
178    #[serde(default)]
179    pub threshold: Option<f64>,
180    /// Minimum file length in lines before a metric rule is evaluated. Below it
181    /// the file is skipped, so ratios are not distorted by tiny files (a small
182    /// module full of required doc comments). `None` means no floor.
183    #[serde(default)]
184    pub min_lines: Option<usize>,
185    /// Project-wide analysis evaluated instead of `rule` or `metric`, and
186    /// mutually exclusive with both. The kind selects builtin Rust logic; the
187    /// YAML only carries the rule's identity, severity and user-facing text.
188    #[serde(default)]
189    pub cross_file: Option<CrossFileKind>,
190    /// Directory a custom rule was loaded from, used to resolve test fixture
191    /// paths. `None` for builtin rules, which resolve against `BuiltinRules`.
192    #[serde(skip)]
193    pub source_dir: Option<PathBuf>,
194    #[serde(default)]
195    pub files: Option<Vec<String>>,
196    #[serde(default)]
197    pub ignores: Option<Vec<String>>,
198    /// Drop findings located inside `#[cfg(test)]` blocks of scanned Rust files.
199    /// Complements `ignores`, which only filters whole files by path.
200    #[serde(default = "default_enabled")]
201    pub enabled: bool,
202    #[serde(default)]
203    pub skip_test_code: bool,
204    #[serde(default)]
205    pub tests: Option<RuleTests>,
206    /// When set, matches of `rule` are candidates confirmed by an LLM before
207    /// being reported. See [`AiCheck`]. The AST scanner ignores this field.
208    #[serde(default)]
209    pub ai_check: Option<AiCheck>,
210}
211
212impl Rule {
213    /// The rule's category, resolving an unset category to the default
214    /// (`Correctness`), the way the scanner treats it.
215    pub fn effective_category(&self) -> Category {
216        self.category.clone().unwrap_or_default()
217    }
218
219    /// Whether this rule measures a file-level metric instead of matching AST.
220    pub fn is_metric(&self) -> bool {
221        self.metric.is_some()
222    }
223
224    /// The metric and threshold of a file-level rule, `None` for AST rules.
225    pub fn metric_spec(&self) -> Option<(Metric, f64)> {
226        Some((self.metric?, self.threshold?))
227    }
228
229    /// Whether this rule is evaluated by the project-wide pass instead of per file.
230    pub fn is_cross_file(&self) -> bool {
231        self.cross_file.is_some()
232    }
233
234    /// The cross-file analysis this rule requests, `None` for AST and metric rules.
235    // Read-only accessor for a private field.
236    // slopguard-disable-next-line no-trivial-function
237    pub fn cross_file_kind(&self) -> Option<CrossFileKind> {
238        self.cross_file
239    }
240}
241
242#[derive(Debug, Error)]
243pub enum RuleError {
244    #[error("failed to parse rule YAML: {0}")]
245    Parse(#[from] serde_yaml::Error),
246
247    #[error("rule '{id}': rule field must not be empty")]
248    EmptyRule { id: RuleId },
249
250    #[error("rule '{id}': 'metric' and 'rule' are mutually exclusive")]
251    MetricAndRule { id: RuleId },
252
253    #[error("rule '{id}': 'cross_file' and 'rule' are mutually exclusive")]
254    CrossFileAndRule { id: RuleId },
255
256    #[error("rule '{id}': 'cross_file' and 'metric' are mutually exclusive")]
257    CrossFileAndMetric { id: RuleId },
258
259    #[error("rule '{id}': a metric rule requires a 'threshold'")]
260    MissingThreshold { id: RuleId },
261
262    #[error("rule '{id}': test fixture files are only supported for metric rules")]
263    FixturesOnAstRule { id: RuleId },
264
265    #[error("rule '{id}': fixture path '{path}' must be relative and must not contain '..'")]
266    InvalidFixturePath { id: RuleId, path: String },
267
268    #[error("rule '{id}': fixture '{path}' not found")]
269    FixtureNotFound { id: RuleId, path: String },
270
271    #[error("duplicate rule id: '{id}'")]
272    DuplicateId { id: RuleId },
273
274    #[error("failed to read rule file '{path}': {source}")]
275    Io {
276        path: String,
277        #[source]
278        source: IoError,
279    },
280
281    #[error("rule file '{path}' is not valid UTF-8: {source}")]
282    Utf8 {
283        path: String,
284        #[source]
285        source: Utf8Error,
286    },
287}
288
289impl RuleError {
290    fn io(path: &Path, source: IoError) -> Self {
291        Self::Io {
292            path: path.display().to_string(),
293            source,
294        }
295    }
296}
297
298/// Parse a single YAML string into a Rule, validating required fields.
299///
300/// A rule carries exactly one matcher: an ast-grep `rule`, a file-level
301/// `metric` plus its `threshold`, or a project-wide `cross_file` kind.
302pub fn parse_rule(yaml: &str) -> Result<Rule, RuleError> {
303    let rule: Rule = serde_yaml::from_str(yaml)?;
304    let has_rule = !rule.rule.is_null();
305    let has_metric = rule.metric.is_some();
306    let has_cross = rule.cross_file.is_some();
307    match (has_cross, has_metric, has_rule) {
308        (true, _, true) => return Err(RuleError::CrossFileAndRule { id: rule.id }),
309        (true, true, _) => return Err(RuleError::CrossFileAndMetric { id: rule.id }),
310        (false, true, true) => return Err(RuleError::MetricAndRule { id: rule.id }),
311        (false, true, false) if rule.threshold.is_none() => {
312            return Err(RuleError::MissingThreshold { id: rule.id })
313        }
314        (false, false, false) => return Err(RuleError::EmptyRule { id: rule.id }),
315        _ => {}
316    }
317    if !rule.is_metric() {
318        if let Some(tests) = &rule.tests {
319            if !tests.should_match_files.is_empty() || !tests.should_not_match_files.is_empty() {
320                return Err(RuleError::FixturesOnAstRule { id: rule.id });
321            }
322        }
323    }
324    Ok(rule)
325}
326
327/// Read a whole-file test fixture referenced by `tests.should_match_files` or
328/// `tests.should_not_match_files`. Builtin rules resolve against the embedded
329/// ruleset; custom rules resolve against the directory they were loaded from.
330pub fn read_fixture(rule: &Rule, relative: &str) -> Result<String, RuleError> {
331    let path = Path::new(relative);
332    let traverses = path.components().any(|c| matches!(c, Component::ParentDir));
333    if path.is_absolute() || traverses {
334        return Err(RuleError::InvalidFixturePath {
335            id: rule.id.clone(),
336            path: relative.to_string(),
337        });
338    }
339
340    match &rule.source_dir {
341        Some(dir) => {
342            let full = dir.join(path);
343            read_to_string(&full).map_err(|e| {
344                if e.kind() == std::io::ErrorKind::NotFound {
345                    RuleError::FixtureNotFound {
346                        id: rule.id.clone(),
347                        path: relative.to_string(),
348                    }
349                } else {
350                    RuleError::io(&full, e)
351                }
352            })
353        }
354        None => {
355            let file = BuiltinRules::get(relative).ok_or_else(|| RuleError::FixtureNotFound {
356                id: rule.id.clone(),
357                path: relative.to_string(),
358            })?;
359            from_utf8(&file.data)
360                .map(str::to_string)
361                .map_err(|e| RuleError::Utf8 {
362                    path: relative.to_string(),
363                    source: e,
364                })
365        }
366    }
367}
368
369/// Parse a rule and fill in its category from the ruleset-relative path when
370/// the YAML omits it.
371fn parse_rule_at(yaml: &str, relative_path: &Path) -> Result<Rule, RuleError> {
372    let mut rule = parse_rule(yaml)?;
373    if rule.category.is_none() {
374        rule.category = Category::from_rule_path(relative_path);
375    }
376    Ok(rule)
377}
378
379fn is_rule_file(path: &Path) -> bool {
380    matches!(
381        path.extension().and_then(|e| e.to_str()),
382        Some("yml" | "yaml")
383    )
384}
385
386/// Load all builtin rules embedded via rust-embed.
387pub fn load_builtin_rules() -> Result<Vec<Rule>, RuleError> {
388    BuiltinRules::iter()
389        .filter(|path| is_rule_file(Path::new(path.as_ref())))
390        .filter_map(|path| BuiltinRules::get(&path).map(|file| (path, file)))
391        .map(|(path, file)| {
392            let yaml = from_utf8(&file.data).map_err(|e| RuleError::Utf8 {
393                path: path.to_string(),
394                source: e,
395            })?;
396            parse_rule_at(yaml, Path::new(path.as_ref()))
397        })
398        .collect()
399}
400
401/// Load rules from custom directories on the filesystem. Missing directories
402/// are skipped; subdirectories are searched recursively.
403pub fn load_custom_rules(dirs: &[PathBuf]) -> Result<Vec<Rule>, RuleError> {
404    let mut rules = Vec::new();
405    for dir in dirs.iter().filter(|dir| dir.is_dir()) {
406        // Custom rule directories are plain file trees: gitignore and
407        // hidden-file filtering would silently drop rules.
408        for entry in WalkBuilder::new(dir).standard_filters(false).build() {
409            let entry = entry.map_err(|e| RuleError::io(dir, IoError::other(e)))?;
410            let path = entry.path();
411            let is_file = entry.file_type().is_some_and(|t| t.is_file());
412            if !is_file || !is_rule_file(path) {
413                continue;
414            }
415            let yaml = read_to_string(path).map_err(|e| RuleError::io(path, e))?;
416            let relative = path.strip_prefix(dir).unwrap_or(path);
417            let mut rule = parse_rule_at(&yaml, relative)?;
418            rule.source_dir = Some(dir.clone());
419            rules.push(rule);
420        }
421    }
422    Ok(rules)
423}
424
425/// Whether a builtin rule takes part in a scan under `config`.
426///
427/// A rule listed in `rules.enable` is always active. Otherwise it must be
428/// enabled by default, belong to an active ruleset, and not be listed in
429/// `rules.disable`.
430pub fn is_rule_active(rule: &Rule, config: &crate::config::Config) -> bool {
431    let listed = |ids: &[String]| ids.iter().any(|id| id == rule.id.as_str());
432    if listed(&config.rules.enable) {
433        return true;
434    }
435    let ruleset_on = match rule.effective_category() {
436        Category::Slop => config.rulesets.slop,
437        Category::Security => config.rulesets.security,
438        Category::Correctness => config.rulesets.correctness,
439    };
440    ruleset_on && rule.enabled && !listed(&config.rules.disable)
441}
442
443/// Load all effective rules based on config: builtin (filtered by rulesets and
444/// the enable/disable lists) plus custom rules from configured directories.
445pub fn load_effective_rules(config: &crate::config::Config) -> Result<Vec<Rule>, RuleError> {
446    let mut rules: Vec<Rule> = load_builtin_rules()?
447        .into_iter()
448        .filter(|rule| is_rule_active(rule, config))
449        .collect();
450
451    let custom = load_custom_rules(&config.rules.custom_dirs)?;
452    rules.extend(custom);
453    validate_unique_ids(&rules)?;
454    Ok(rules)
455}
456
457/// Load all rules (builtin + custom) without filtering by activation status.
458/// Used by `explain` to look up any rule regardless of config.
459pub fn load_all_rules(config: &crate::config::Config) -> Result<Vec<Rule>, RuleError> {
460    let mut rules = load_builtin_rules()?;
461    let custom = load_custom_rules(&config.rules.custom_dirs)?;
462    rules.extend(custom);
463    Ok(rules)
464}
465
466/// Validate that all rule (id, language) pairs are unique. The same rule id
467/// may appear for different languages (e.g. `no-todo-fixme` for both Rust and
468/// TypeScript), so a single `rules.disable` entry covers both variants.
469pub fn validate_unique_ids(rules: &[Rule]) -> Result<(), RuleError> {
470    let mut seen = HashSet::new();
471    for rule in rules {
472        if !seen.insert((&rule.id, &rule.language)) {
473            return Err(RuleError::DuplicateId {
474                id: rule.id.clone(),
475            });
476        }
477    }
478    Ok(())
479}
480
481#[cfg(test)]
482mod load_tests;
483#[cfg(test)]
484mod parse_tests;
slopguard_core/rule/mod.rs

slopguard_core/rule/
mod.rs
