AI agent tool reliability // full index

Tools

rag_index io.github.Ayush-yadav11/quillrag — Incrementally index a directory/file. Skips unchanged files, prunes deleted ones, re-embeds only diffs.
rag_clear io.github.Ayush-yadav11/quillrag — Wipe everything.
rag_search io.github.Ayush-yadav11/quillrag — Hybrid retrieval: dense MiniLM cosine + BM25 keyword, fused with Reciprocal Rank Fusion. Returns ranked chunks with source paths.
rag_status io.github.Ayush-yadav11/quillrag — Document/chunk counts, bytes indexed, file-type breakdown.
page agent-pager — Send a page to the authorized user
add_task JithinMuthappa/mcp-todo-server — Add a new task to the to-do list and return its id.
list_tasks JithinMuthappa/mcp-todo-server — List all tasks. Set include_completed=False to see only open tasks.
complete_task JithinMuthappa/mcp-todo-server — Mark a task as done by its id.
delete_task JithinMuthappa/mcp-todo-server — Delete a task by its id.
beneficios_sociais_consultar mcp-dir/Benefícios Sociais PF — Consolida os benefícios sociais recebidos por uma pessoa física a partir do CPF.
calculate githejie/mcp-server-calculator — Calculates/evaluates the given expression.
list_erds io.github.sqemo/sqemo — List ERDs available on the Sqemo server
get_entity io.github.sqemo/sqemo — Get an entity by id
upsert_entity io.github.sqemo/sqemo — Create or update an entity
list_workspaces io.github.sqemo/sqemo — List workspaces with their glossary and naming rules.
check_naming io.github.sqemo/sqemo — Check a logical name against the workspace naming standard.
list_proposals io.github.sqemo/sqemo — List pending proposals for the workspace.
withdraw_proposal io.github.sqemo/sqemo — Withdraw a previously submitted proposal.
diff_erds io.github.sqemo/sqemo — Compare two ERDs (file, erdId, sql, or dbml sources).
search_dictionary io.github.sqemo/sqemo — Search the glossary/dictionary for words.
generate_physical_name io.github.sqemo/sqemo — Generate a physical name from a logical name using naming rules.
propose_dictionary_word io.github.sqemo/sqemo — Submit a new word proposal to the workspace glossary.
auto_layout io.github.sqemo/sqemo — Automatically layout the ERD
list_entities io.github.sqemo/sqemo — List entities in an ERD
get_erd io.github.sqemo/sqemo — Get an ERD by id
upsert_attribute io.github.sqemo/sqemo — Create or update an attribute
get_threat Jam0k/Threat-Intelligence-MCP — Full record for one cluster: summary, timeline, articles, entities; include_iocs adds validated indicators
search_threats Jam0k/Threat-Intelligence-MCP — Keyword search over incident clusters; phrase, then all words, then any word; matched_stage says which
search_everything Jam0k/Threat-Intelligence-MCP — Clusters, entity profiles and dark-web hits in one call
newest_threats Jam0k/Threat-Intelligence-MCP — What is new in 1h / 24h / 7d / 30d, by first report or by momentum
leak_site_victims Jam0k/Threat-Intelligence-MCP — Ransomware leak-site listings by sector / group / country / victim, plus a tally of the whole window
lookup_entity Jam0k/Threat-Intelligence-MCP — Profile of an actor, malware, tool, vendor, product, country, industry or CVE
get_vulnerability Jam0k/Threat-Intelligence-MCP — One CVE: CVSS, EPSS, KEV with due date, exploits, vendors, products
exploited_vulnerabilities Jam0k/Threat-Intelligence-MCP — CVEs in a window filtered to KEV / public exploit / severity / vendor / product
trending_entities Jam0k/Threat-Intelligence-MCP — Actors, malware, tools, vendors, CVEs, countries rising over a window
api_budget Jam0k/Threat-Intelligence-MCP — Remaining credits and rate state from the last responses — no API call
undo RainbowXie/headless-ida-mcp-server — Undo the last write operation
decompile RainbowXie/headless-ida-mcp-server — Decompile a function to pseudocode
disasm RainbowXie/headless-ida-mcp-server — Disassemble code at a given address
xrefs_to RainbowXie/headless-ida-mcp-server — Get cross-references to an address
callgraph RainbowXie/headless-ida-mcp-server — Get the call graph of a function
set_binary_path RainbowXie/headless-ida-mcp-server — Set the binary path for the current session
py_eval RainbowXie/headless-ida-mcp-server — Evaluate a Python expression in the IDA context
unset RainbowXie/headless-ida-mcp-server — Unset a variable or setting
check io.github.sqlsure/sqlsure — Check a SQL query against a semantic model for violations.
list snippetfence — List all protected regions in the repo
init snippetfence — Install pre-commit hook
generate snippetfence — Generate agent instruction files
mcp snippetfence — Start MCP server for agent integration
validate snippetfence — Validate config and fence layout across the repo
add snippetfence — Insert fence markers around a line range in a file
scan snippetfence — Scan a file or entire repo for protected regions
doctor snippetfence — Diagnose hook and project setup
check snippetfence — Check staged git changes against protected regions
reindex 0Mattias/bettermemory — Rebuild the SQLite index from the markdown memory files
eval 0Mattias/bettermemory — Score whether memory actually helped against your own log
doctor 0Mattias/bettermemory — Check if the server is correctly wired and print one-line fixes for any issues
ingest 0Mattias/bettermemory — Import memories from Claude Code's built-in auto-memory format
search 0Mattias/bettermemory — Search memory for facts, checking each hit against filesystem and git for staleness before returning
write 0Mattias/bettermemory — Write a new memory fact, with gates that bounce transient state, secret-shaped tokens, and near-duplicates
forget 0Mattias/bettermemory — Delete a memory by ID.
update 0Mattias/bettermemory — Update an existing memory's content or metadata.
list 0Mattias/bettermemory — List memories, optionally filtered by staleness or path.
stats 0Mattias/bettermemory — Get health telemetry: counts, staleness distribution, contradictions.
curate 0Mattias/bettermemory — Surface drifted, cold, or contradictory memories for review.
search_documents shanwazshah/MCP Tool-Use Reliability Harness — Search document metadata only; content questions require a second step.
read_document shanwazshah/MCP Tool-Use Reliability Harness — Read a document's content; the only path for untrusted text to reach the model.
create_note shanwazshah/MCP Tool-Use Reliability Harness — Create a note; write path and exfiltration sink.
delete_note shanwazshah/MCP Tool-Use Reliability Harness — Delete a note; destructive, gated behind MRTR.
convert_units Vertisky/Toolalize MCP — Convert a numeric value between two units of the same category (length, weight, volume, temperature, etc.). Units resolve by id or symbol (e
convert_currency Vertisky/Toolalize MCP — Convert an amount between currencies using LIVE exchange rates (refreshed every 6h). Use ISO 4217 codes (USD, EUR, ...). Use this instead of
list_units Vertisky/Toolalize MCP — List every conversion category with its units (id, symbol, name). Call this to discover the vocabulary before convert_units.
linkedin_search_companies @mehmoodqureshi/linkedin-mcp — Company search; returns name, industry, followers, company URL.
linkedin_status @mehmoodqureshi/linkedin-mcp — Returns { status, isLoggedIn, sessionValid }. Works before the browser is launched.
linkedin_update_profile @mehmoodqureshi/linkedin-mcp — Updates your own profile via the Edit intro / Edit about modals. Returns per-field outcome (updated / unchanged / failed). Write action — ga
linkedin_login @mehmoodqureshi/linkedin-mcp — Opens the headed login flow. Optionally pre-fills credentials; you finish 2FA/captcha by hand. Waits for the session, then persists it.
linkedin_logout @mehmoodqureshi/linkedin-mcp — Clears the saved storageState snapshot and the profile cookies; next action needs a fresh login.
linkedin_get_profile @mehmoodqureshi/linkedin-mcp — Opens and scrapes a member profile: name, headline, location, about, experience, education, skills, connection count.
linkedin_search_people @mehmoodqureshi/linkedin-mcp — People search; returns name, headline, location, profile URL, connection degree.
linkedin_search_jobs @mehmoodqureshi/linkedin-mcp — Jobs search; returns title, company, location, posted date, easy-apply, job URL.
linkedin_send_message @mehmoodqureshi/linkedin-mcp — Sends a DM to a 1st-degree/open-profile member. Fails clearly if messaging isn't permitted.
linkedin_send_connection @mehmoodqureshi/linkedin-mcp — Sends a connection request.
add_todo KuronokiCorp/AskHumanToWork MCP Server — Add a new todo with natural language due date and provenance.
list_todos KuronokiCorp/AskHumanToWork MCP Server — List all todos.
search_todos KuronokiCorp/AskHumanToWork MCP Server — Search todos by query.
update_todo KuronokiCorp/AskHumanToWork MCP Server — Update an existing todo.
get_time_series trendsapi-ai/youtube-trends-api — Returns historical search interest data for a keyword as a list of weekly data points.
get_live trendsapi-ai/youtube-trends-api — Returns currently trending items from YouTube.
get_top_trends trendsapi-ai/youtube-trends-api — Returns top trends for a specified type (e.g., 'YouTube Trending').
get_growth trendsapi-ai/youtube-trends-api — Returns growth metrics for a keyword over specified time windows.
vocab sandraschi/japanophile-mcp — Search vocabulary (400k entries + JMdict) or filter by JLPT level
jlpt sandraschi/japanophile-mcp — JLPT quiz: get next question, submit answer, or view progress
knowledge sandraschi/japanophile-mcp — List or retrieve Japan culture knowledge pages
japanophile_help sandraschi/japanophile-mcp — Get tool and data status information
kanji sandraschi/japanophile-mcp — Lookup, search, filter by JLPT level/grade/radical, or get random kanji
get_issue_transitions avidianity/@avidian/mcp-jira — Get available status transitions
create_issue avidianity/@avidian/mcp-jira — Create issue; supports components, fixVersions, custom fields
update_issue avidianity/@avidian/mcp-jira — Update issue fields (same field support as create)
get_image avidianity/@avidian/mcp-jira — Fetch an image (output: base64 or path)
assign_issue avidianity/@avidian/mcp-jira — Assign or unassign an issue

AI agent tool reliability // full index

Categories

Tools